1 // © 2016 and later: Unicode, Inc. and others.
2 // License & terms of use: http://www.unicode.org/copyright.html
3 /*
4 *******************************************************************************
5 * Copyright (C) 1997-2015, International Business Machines Corporation and others.
6 * All Rights Reserved.
7 *******************************************************************************
8 */
9
10 #ifndef RBNF_H
11 #define RBNF_H
12
13 #include "unicode/utypes.h"
14
15 /**
16 * \file
17 * \brief C++ API: Rule Based Number Format
18 */
19
20 /**
21 * \def U_HAVE_RBNF
22 * This will be 0 if RBNF support is not included in ICU
23 * and 1 if it is.
24 *
25 * @stable ICU 2.4
26 */
27 #if UCONFIG_NO_FORMATTING
28 #define U_HAVE_RBNF 0
29 #else
30 #define U_HAVE_RBNF 1
31
32 #include "unicode/dcfmtsym.h"
33 #include "unicode/fmtable.h"
34 #include "unicode/locid.h"
35 #include "unicode/numfmt.h"
36 #include "unicode/unistr.h"
37 #include "unicode/strenum.h"
38 #include "unicode/brkiter.h"
39 #include "unicode/upluralrules.h"
40
41 U_NAMESPACE_BEGIN
42
43 class NFRule;
44 class NFRuleSet;
45 class LocalizationInfo;
46 class PluralFormat;
47 class RuleBasedCollator;
48
49 /**
50 * Tags for the predefined rulesets.
51 *
52 * @stable ICU 2.2
53 */
54 enum URBNFRuleSetTag {
55 URBNF_SPELLOUT,
56 URBNF_ORDINAL,
57 URBNF_DURATION,
58 URBNF_NUMBERING_SYSTEM,
59 #ifndef U_HIDE_DEPRECATED_API
60 /**
61 * One more than the highest normal URBNFRuleSetTag value.
62 * @deprecated ICU 58 The numeric value may change over time, see ICU ticket #12420.
63 */
64 URBNF_COUNT
65 #endif // U_HIDE_DEPRECATED_API
66 };
67
68 /**
69 * The RuleBasedNumberFormat class formats numbers according to a set of rules. This number formatter is
70 * typically used for spelling out numeric values in words (e.g., 25,3476 as
71 * "twenty-five thousand three hundred seventy-six" or "vingt-cinq mille trois
72 * cents soixante-seize" or
73 * "fünfundzwanzigtausenddreihundertsechsundsiebzig"), but can also be used for
74 * other complicated formatting tasks, such as formatting a number of seconds as hours,
75 * minutes and seconds (e.g., 3,730 as "1:02:10").
76 *
77 * <p>The resources contain three predefined formatters for each locale: spellout, which
78 * spells out a value in words (123 is "one hundred twenty-three"); ordinal, which
79 * appends an ordinal suffix to the end of a numeral (123 is "123rd"); and
80 * duration, which shows a duration in seconds as hours, minutes, and seconds (123 is
81 * "2:03"). The client can also define more specialized <tt>RuleBasedNumberFormat</tt>s
82 * by supplying programmer-defined rule sets.</p>
83 *
84 * <p>The behavior of a <tt>RuleBasedNumberFormat</tt> is specified by a textual description
85 * that is either passed to the constructor as a <tt>String</tt> or loaded from a resource
86 * bundle. In its simplest form, the description consists of a semicolon-delimited list of <em>rules.</em>
87 * Each rule has a string of output text and a value or range of values it is applicable to.
88 * In a typical spellout rule set, the first twenty rules are the words for the numbers from
89 * 0 to 19:</p>
90 *
91 * <pre>zero; one; two; three; four; five; six; seven; eight; nine;
92 * ten; eleven; twelve; thirteen; fourteen; fifteen; sixteen; seventeen; eighteen; nineteen;</pre>
93 *
94 * <p>For larger numbers, we can use the preceding set of rules to format the ones place, and
95 * we only have to supply the words for the multiples of 10:</p>
96 *
97 * <pre> 20: twenty[->>];
98 * 30: thirty[->>];
99 * 40: forty[->>];
100 * 50: fifty[->>];
101 * 60: sixty[->>];
102 * 70: seventy[->>];
103 * 80: eighty[->>];
104 * 90: ninety[->>];</pre>
105 *
106 * <p>In these rules, the <em>base value</em> is spelled out explicitly and set off from the
107 * rule's output text with a colon. The rules are in a sorted list, and a rule is applicable
108 * to all numbers from its own base value to one less than the next rule's base value. The
109 * ">>" token is called a <em>substitution</em> and tells the fomatter to
110 * isolate the number's ones digit, format it using this same set of rules, and place the
111 * result at the position of the ">>" token. Text in brackets is omitted if
112 * the number being formatted is an even multiple of 10 (the hyphen is a literal hyphen; 24
113 * is "twenty-four," not "twenty four").</p>
114 *
115 * <p>For even larger numbers, we can actually look up several parts of the number in the
116 * list:</p>
117 *
118 * <pre>100: << hundred[ >>];</pre>
119 *
120 * <p>The "<<" represents a new kind of substitution. The << isolates
121 * the hundreds digit (and any digits to its left), formats it using this same rule set, and
122 * places the result where the "<<" was. Notice also that the meaning of
123 * >> has changed: it now refers to both the tens and the ones digits. The meaning of
124 * both substitutions depends on the rule's base value. The base value determines the rule's <em>divisor,</em>
125 * which is the highest power of 10 that is less than or equal to the base value (the user
126 * can change this). To fill in the substitutions, the formatter divides the number being
127 * formatted by the divisor. The integral quotient is used to fill in the <<
128 * substitution, and the remainder is used to fill in the >> substitution. The meaning
129 * of the brackets changes similarly: text in brackets is omitted if the value being
130 * formatted is an even multiple of the rule's divisor. The rules are applied recursively, so
131 * if a substitution is filled in with text that includes another substitution, that
132 * substitution is also filled in.</p>
133 *
134 * <p>This rule covers values up to 999, at which point we add another rule:</p>
135 *
136 * <pre>1000: << thousand[ >>];</pre>
137 *
138 * <p>Again, the meanings of the brackets and substitution tokens shift because the rule's
139 * base value is a higher power of 10, changing the rule's divisor. This rule can actually be
140 * used all the way up to 999,999. This allows us to finish out the rules as follows:</p>
141 *
142 * <pre> 1,000,000: << million[ >>];
143 * 1,000,000,000: << billion[ >>];
144 * 1,000,000,000,000: << trillion[ >>];
145 * 1,000,000,000,000,000: OUT OF RANGE!;</pre>
146 *
147 * <p>Commas, periods, and spaces can be used in the base values to improve legibility and
148 * are ignored by the rule parser. The last rule in the list is customarily treated as an
149 * "overflow rule," applying to everything from its base value on up, and often (as
150 * in this example) being used to print out an error message or default representation.
151 * Notice also that the size of the major groupings in large numbers is controlled by the
152 * spacing of the rules: because in English we group numbers by thousand, the higher rules
153 * are separated from each other by a factor of 1,000.</p>
154 *
155 * <p>To see how these rules actually work in practice, consider the following example:
156 * Formatting 25,430 with this rule set would work like this:</p>
157 *
158 * <table border="0" width="100%">
159 * <tr>
160 * <td><strong><< thousand >></strong></td>
161 * <td>[the rule whose base value is 1,000 is applicable to 25,340]</td>
162 * </tr>
163 * <tr>
164 * <td><strong>twenty->></strong> thousand >></td>
165 * <td>[25,340 over 1,000 is 25. The rule for 20 applies.]</td>
166 * </tr>
167 * <tr>
168 * <td>twenty-<strong>five</strong> thousand >></td>
169 * <td>[25 mod 10 is 5. The rule for 5 is "five."</td>
170 * </tr>
171 * <tr>
172 * <td>twenty-five thousand <strong><< hundred >></strong></td>
173 * <td>[25,340 mod 1,000 is 340. The rule for 100 applies.]</td>
174 * </tr>
175 * <tr>
176 * <td>twenty-five thousand <strong>three</strong> hundred >></td>
177 * <td>[340 over 100 is 3. The rule for 3 is "three."]</td>
178 * </tr>
179 * <tr>
180 * <td>twenty-five thousand three hundred <strong>forty</strong></td>
181 * <td>[340 mod 100 is 40. The rule for 40 applies. Since 40 divides
182 * evenly by 10, the hyphen and substitution in the brackets are omitted.]</td>
183 * </tr>
184 * </table>
185 *
186 * <p>The above syntax suffices only to format positive integers. To format negative numbers,
187 * we add a special rule:</p>
188 *
189 * <pre>-x: minus >>;</pre>
190 *
191 * <p>This is called a <em>negative-number rule,</em> and is identified by "-x"
192 * where the base value would be. This rule is used to format all negative numbers. the
193 * >> token here means "find the number's absolute value, format it with these
194 * rules, and put the result here."</p>
195 *
196 * <p>We also add a special rule called a <em>fraction rule </em>for numbers with fractional
197 * parts:</p>
198 *
199 * <pre>x.x: << point >>;</pre>
200 *
201 * <p>This rule is used for all positive non-integers (negative non-integers pass through the
202 * negative-number rule first and then through this rule). Here, the << token refers to
203 * the number's integral part, and the >> to the number's fractional part. The
204 * fractional part is formatted as a series of single-digit numbers (e.g., 123.456 would be
205 * formatted as "one hundred twenty-three point four five six").</p>
206 *
207 * <p>To see how this rule syntax is applied to various languages, examine the resource data.</p>
208 *
209 * <p>There is actually much more flexibility built into the rule language than the
210 * description above shows. A formatter may own multiple rule sets, which can be selected by
211 * the caller, and which can use each other to fill in their substitutions. Substitutions can
212 * also be filled in with digits, using a DecimalFormat object. There is syntax that can be
213 * used to alter a rule's divisor in various ways. And there is provision for much more
214 * flexible fraction handling. A complete description of the rule syntax follows:</p>
215 *
216 * <hr>
217 *
218 * <p>The description of a <tt>RuleBasedNumberFormat</tt>'s behavior consists of one or more <em>rule
219 * sets.</em> Each rule set consists of a name, a colon, and a list of <em>rules.</em> A rule
220 * set name must begin with a % sign. Rule sets with names that begin with a single % sign
221 * are <em>public:</em> the caller can specify that they be used to format and parse numbers.
222 * Rule sets with names that begin with %% are <em>private:</em> they exist only for the use
223 * of other rule sets. If a formatter only has one rule set, the name may be omitted.</p>
224 *
225 * <p>The user can also specify a special "rule set" named <tt>%%lenient-parse</tt>.
226 * The body of <tt>%%lenient-parse</tt> isn't a set of number-formatting rules, but a <tt>RuleBasedCollator</tt>
227 * description which is used to define equivalences for lenient parsing. For more information
228 * on the syntax, see <tt>RuleBasedCollator</tt>. For more information on lenient parsing,
229 * see <tt>setLenientParse()</tt>. <em>Note:</em> symbols that have syntactic meaning
230 * in collation rules, such as '&', have no particular meaning when appearing outside
231 * of the <tt>lenient-parse</tt> rule set.</p>
232 *
233 * <p>The body of a rule set consists of an ordered, semicolon-delimited list of <em>rules.</em>
234 * Internally, every rule has a base value, a divisor, rule text, and zero, one, or two <em>substitutions.</em>
235 * These parameters are controlled by the description syntax, which consists of a <em>rule
236 * descriptor,</em> a colon, and a <em>rule body.</em></p>
237 *
238 * <p>A rule descriptor can take one of the following forms (text in <em>italics</em> is the
239 * name of a token):</p>
240 *
241 * <table border="0" width="100%">
242 * <tr>
243 * <td><em>bv</em>:</td>
244 * <td><em>bv</em> specifies the rule's base value. <em>bv</em> is a decimal
245 * number expressed using ASCII digits. <em>bv</em> may contain spaces, period, and commas,
246 * which are ignored. The rule's divisor is the highest power of 10 less than or equal to
247 * the base value.</td>
248 * </tr>
249 * <tr>
250 * <td><em>bv</em>/<em>rad</em>:</td>
251 * <td><em>bv</em> specifies the rule's base value. The rule's divisor is the
252 * highest power of <em>rad</em> less than or equal to the base value.</td>
253 * </tr>
254 * <tr>
255 * <td><em>bv</em>>:</td>
256 * <td><em>bv</em> specifies the rule's base value. To calculate the divisor,
257 * let the radix be 10, and the exponent be the highest exponent of the radix that yields a
258 * result less than or equal to the base value. Every > character after the base value
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><em>bv</em>/<em>rad</em>>:</td>
264 * <td><em>bv</em> specifies the rule's base value. To calculate the divisor,
265 * let the radix be <em>rad</em>, and the exponent be the highest exponent of the radix that
266 * yields a result less than or equal to the base value. Every > character after the radix
267 * decreases the exponent by 1. If the exponent is positive or 0, the divisor is the radix
268 * raised to the power of the exponent; otherwise, the divisor is 1.</td>
269 * </tr>
270 * <tr>
271 * <td>-x:</td>
272 * <td>The rule is a negative-number rule.</td>
273 * </tr>
274 * <tr>
275 * <td>x.x:</td>
276 * <td>The rule is an <em>improper fraction rule</em>. If the full stop in
277 * the middle of the rule name is replaced with the decimal point
278 * that is used in the language or DecimalFormatSymbols, then that rule will
279 * have precedence when formatting and parsing this rule. For example, some
280 * languages use the comma, and can thus be written as x,x instead. For example,
281 * you can use "x.x: << point >>;x,x: << comma >>;" to
282 * handle the decimal point that matches the language's natural spelling of
283 * the punctuation of either the full stop or comma.</td>
284 * </tr>
285 * <tr>
286 * <td>0.x:</td>
287 * <td>The rule is a <em>proper fraction rule</em>. If the full stop in
288 * the middle of the rule name is replaced with the decimal point
289 * that is used in the language or DecimalFormatSymbols, then that rule will
290 * have precedence when formatting and parsing this rule. For example, some
291 * languages use the comma, and can thus be written as 0,x instead. For example,
292 * you can use "0.x: point >>;0,x: comma >>;" to
293 * handle the decimal point that matches the language's natural spelling of
294 * the punctuation of either the full stop or comma.</td>
295 * </tr>
296 * <tr>
297 * <td>x.0:</td>
298 * <td>The rule is a <em>master rule</em>. If the full stop in
299 * the middle of the rule name is replaced with the decimal point
300 * that is used in the language or DecimalFormatSymbols, then that rule will
301 * have precedence when formatting and parsing this rule. For example, some
302 * languages use the comma, and can thus be written as x,0 instead. For example,
303 * you can use "x.0: << point;x,0: << comma;" to
304 * handle the decimal point that matches the language's natural spelling of
305 * the punctuation of either the full stop or comma.</td>
306 * </tr>
307 * <tr>
308 * <td>Inf:</td>
309 * <td>The rule for infinity.</td>
310 * </tr>
311 * <tr>
312 * <td>NaN:</td>
313 * <td>The rule for an IEEE 754 NaN (not a number).</td>
314 * </tr>
315 * <tr>
316 * <td><em>nothing</em></td>
317 * <td>If the rule's rule descriptor is left out, the base value is one plus the
318 * preceding rule's base value (or zero if this is the first rule in the list) in a normal
319 * rule set. In a fraction rule set, the base value is the same as the preceding rule's
320 * base value.</td>
321 * </tr>
322 * </table>
323 *
324 * <p>A rule set may be either a regular rule set or a <em>fraction rule set,</em> depending
325 * on whether it is used to format a number's integral part (or the whole number) or a
326 * number's fractional part. Using a rule set to format a rule's fractional part makes it a
327 * fraction rule set.</p>
328 *
329 * <p>Which rule is used to format a number is defined according to one of the following
330 * algorithms: If the rule set is a regular rule set, do the following:
331 *
332 * <ul>
333 * <li>If the rule set includes a master rule (and the number was passed in as a <tt>double</tt>),
334 * use the master rule. (If the number being formatted was passed in as a <tt>long</tt>,
335 * the master rule is ignored.)</li>
336 * <li>If the number is negative, use the negative-number rule.</li>
337 * <li>If the number has a fractional part and is greater than 1, use the improper fraction
338 * rule.</li>
339 * <li>If the number has a fractional part and is between 0 and 1, use the proper fraction
340 * rule.</li>
341 * <li>Binary-search the rule list for the rule with the highest base value less than or equal
342 * to the number. If that rule has two substitutions, its base value is not an even multiple
343 * of its divisor, and the number <em>is</em> an even multiple of the rule's divisor, use the
344 * rule that precedes it in the rule list. Otherwise, use the rule itself.</li>
345 * </ul>
346 *
347 * <p>If the rule set is a fraction rule set, do the following:
348 *
349 * <ul>
350 * <li>Ignore negative-number and fraction rules.</li>
351 * <li>For each rule in the list, multiply the number being formatted (which will always be
352 * between 0 and 1) by the rule's base value. Keep track of the distance between the result
353 * the nearest integer.</li>
354 * <li>Use the rule that produced the result closest to zero in the above calculation. In the
355 * event of a tie or a direct hit, use the first matching rule encountered. (The idea here is
356 * to try each rule's base value as a possible denominator of a fraction. Whichever
357 * denominator produces the fraction closest in value to the number being formatted wins.) If
358 * the rule following the matching rule has the same base value, use it if the numerator of
359 * the fraction is anything other than 1; if the numerator is 1, use the original matching
360 * rule. (This is to allow singular and plural forms of the rule text without a lot of extra
361 * hassle.)</li>
362 * </ul>
363 *
364 * <p>A rule's body consists of a string of characters terminated by a semicolon. The rule
365 * may include zero, one, or two <em>substitution tokens,</em> and a range of text in
366 * brackets. The brackets denote optional text (and may also include one or both
367 * substitutions). The exact meanings of the substitution tokens, and under what conditions
368 * optional text is omitted, depend on the syntax of the substitution token and the context.
369 * The rest of the text in a rule body is literal text that is output when the rule matches
370 * the number being formatted.</p>
371 *
372 * <p>A substitution token begins and ends with a <em>token character.</em> The token
373 * character and the context together specify a mathematical operation to be performed on the
374 * number being formatted. An optional <em>substitution descriptor </em>specifies how the
375 * value resulting from that operation is used to fill in the substitution. The position of
376 * the substitution token in the rule body specifies the location of the resultant text in
377 * the original rule text.</p>
378 *
379 * <p>The meanings of the substitution token characters are as follows:</p>
380 *
381 * <table border="0" width="100%">
382 * <tr>
383 * <td>>></td>
384 * <td>in normal rule</td>
385 * <td>Divide the number by the rule's divisor and format the remainder</td>
386 * </tr>
387 * <tr>
388 * <td></td>
389 * <td>in negative-number rule</td>
390 * <td>Find the absolute value of the number and format the result</td>
391 * </tr>
392 * <tr>
393 * <td></td>
394 * <td>in fraction or master rule</td>
395 * <td>Isolate the number's fractional part and format it.</td>
396 * </tr>
397 * <tr>
398 * <td></td>
399 * <td>in rule in fraction rule set</td>
400 * <td>Not allowed.</td>
401 * </tr>
402 * <tr>
403 * <td>>>></td>
404 * <td>in normal rule</td>
405 * <td>Divide the number by the rule's divisor and format the remainder,
406 * but bypass the normal rule-selection process and just use the
407 * rule that precedes this one in this rule list.</td>
408 * </tr>
409 * <tr>
410 * <td></td>
411 * <td>in all other rules</td>
412 * <td>Not allowed.</td>
413 * </tr>
414 * <tr>
415 * <td><<</td>
416 * <td>in normal rule</td>
417 * <td>Divide the number by the rule's divisor and format the quotient</td>
418 * </tr>
419 * <tr>
420 * <td></td>
421 * <td>in negative-number rule</td>
422 * <td>Not allowed.</td>
423 * </tr>
424 * <tr>
425 * <td></td>
426 * <td>in fraction or master rule</td>
427 * <td>Isolate the number's integral part and format it.</td>
428 * </tr>
429 * <tr>
430 * <td></td>
431 * <td>in rule in fraction rule set</td>
432 * <td>Multiply the number by the rule's base value and format the result.</td>
433 * </tr>
434 * <tr>
435 * <td>==</td>
436 * <td>in all rule sets</td>
437 * <td>Format the number unchanged</td>
438 * </tr>
439 * <tr>
440 * <td>[]</td>
441 * <td>in normal rule</td>
442 * <td>Omit the optional text if the number is an even multiple of the rule's divisor</td>
443 * </tr>
444 * <tr>
445 * <td></td>
446 * <td>in negative-number rule</td>
447 * <td>Not allowed.</td>
448 * </tr>
449 * <tr>
450 * <td></td>
451 * <td>in improper-fraction rule</td>
452 * <td>Omit the optional text if the number is between 0 and 1 (same as specifying both an
453 * x.x rule and a 0.x rule)</td>
454 * </tr>
455 * <tr>
456 * <td></td>
457 * <td>in master rule</td>
458 * <td>Omit the optional text if the number is an integer (same as specifying both an x.x
459 * rule and an x.0 rule)</td>
460 * </tr>
461 * <tr>
462 * <td></td>
463 * <td>in proper-fraction rule</td>
464 * <td>Not allowed.</td>
465 * </tr>
466 * <tr>
467 * <td></td>
468 * <td>in rule in fraction rule set</td>
469 * <td>Omit the optional text if multiplying the number by the rule's base value yields 1.</td>
470 * </tr>
471 * <tr>
472 * <td width="37">$(cardinal,<i>plural syntax</i>)$</td>
473 * <td width="23"></td>
474 * <td width="165" valign="top">in all rule sets</td>
475 * <td>This provides the ability to choose a word based on the number divided by the radix to the power of the
476 * exponent of the base value for the specified locale, which is normally equivalent to the << value.
477 * This uses the cardinal plural rules from PluralFormat. All strings used in the plural format are treated
478 * as the same base value for parsing.</td>
479 * </tr>
480 * <tr>
481 * <td width="37">$(ordinal,<i>plural syntax</i>)$</td>
482 * <td width="23"></td>
483 * <td width="165" valign="top">in all rule sets</td>
484 * <td>This provides the ability to choose a word based on the number divided by the radix to the power of the
485 * exponent of the base value for the specified locale, which is normally equivalent to the << value.
486 * This uses the ordinal plural rules from PluralFormat. All strings used in the plural format are treated
487 * as the same base value for parsing.</td>
488 * </tr>
489 * </table>
490 *
491 * <p>The substitution descriptor (i.e., the text between the token characters) may take one
492 * of three forms:</p>
493 *
494 * <table border="0" width="100%">
495 * <tr>
496 * <td>a rule set name</td>
497 * <td>Perform the mathematical operation on the number, and format the result using the
498 * named rule set.</td>
499 * </tr>
500 * <tr>
501 * <td>a DecimalFormat pattern</td>
502 * <td>Perform the mathematical operation on the number, and format the result using a
503 * DecimalFormat with the specified pattern. The pattern must begin with 0 or #.</td>
504 * </tr>
505 * <tr>
506 * <td>nothing</td>
507 * <td>Perform the mathematical operation on the number, and format the result using the rule
508 * set containing the current rule, except:
509 * <ul>
510 * <li>You can't have an empty substitution descriptor with a == substitution.</li>
511 * <li>If you omit the substitution descriptor in a >> substitution in a fraction rule,
512 * format the result one digit at a time using the rule set containing the current rule.</li>
513 * <li>If you omit the substitution descriptor in a << substitution in a rule in a
514 * fraction rule set, format the result using the default rule set for this formatter.</li>
515 * </ul>
516 * </td>
517 * </tr>
518 * </table>
519 *
520 * <p>Whitespace is ignored between a rule set name and a rule set body, between a rule
521 * descriptor and a rule body, or between rules. If a rule body begins with an apostrophe,
522 * the apostrophe is ignored, but all text after it becomes significant (this is how you can
523 * have a rule's rule text begin with whitespace). There is no escape function: the semicolon
524 * is not allowed in rule set names or in rule text, and the colon is not allowed in rule set
525 * names. The characters beginning a substitution token are always treated as the beginning
526 * of a substitution token.</p>
527 *
528 * <p>See the resource data and the demo program for annotated examples of real rule sets
529 * using these features.</p>
530 *
531 * <p><em>User subclasses are not supported.</em> While clients may write
532 * subclasses, such code will not necessarily work and will not be
533 * guaranteed to work stably from release to release.
534 *
535 * <p><b>Localizations</b></p>
536 * <p>Constructors are available that allow the specification of localizations for the
537 * public rule sets (and also allow more control over what public rule sets are available).
538 * Localization data is represented as a textual description. The description represents
539 * an array of arrays of string. The first element is an array of the public rule set names,
540 * each of these must be one of the public rule set names that appear in the rules. Only
541 * names in this array will be treated as public rule set names by the API. Each subsequent
542 * element is an array of localizations of these names. The first element of one of these
543 * subarrays is the locale name, and the remaining elements are localizations of the
544 * public rule set names, in the same order as they were listed in the first arrray.</p>
545 * <p>In the syntax, angle brackets '<', '>' are used to delimit the arrays, and comma ',' is used
546 * to separate elements of an array. Whitespace is ignored, unless quoted.</p>
547 * <p>For example:<pre>
548 * < < %foo, %bar, %baz >,
549 * < en, Foo, Bar, Baz >,
550 * < fr, 'le Foo', 'le Bar', 'le Baz' >
551 * < zh, \\u7532, \\u4e59, \\u4e19 > >
552 * </pre></p>
553 * @author Richard Gillam
554 * @see NumberFormat
555 * @see DecimalFormat
556 * @see PluralFormat
557 * @see PluralRules
558 * @stable ICU 2.0
559 */
560 class U_I18N_API RuleBasedNumberFormat : public NumberFormat {
561 public:
562
563 //-----------------------------------------------------------------------
564 // constructors
565 //-----------------------------------------------------------------------
566
567 /**
568 * Creates a RuleBasedNumberFormat that behaves according to the description
569 * passed in. The formatter uses the default locale.
570 * @param rules A description of the formatter's desired behavior.
571 * See the class documentation for a complete explanation of the description
572 * syntax.
573 * @param perror The parse error if an error was encountered.
574 * @param status The status indicating whether the constructor succeeded.
575 * @stable ICU 3.2
576 */
577 RuleBasedNumberFormat(const UnicodeString& rules, UParseError& perror, UErrorCode& status);
578
579 /**
580 * Creates a RuleBasedNumberFormat that behaves according to the description
581 * passed in. The formatter uses the default locale.
582 * <p>
583 * The localizations data provides information about the public
584 * rule sets and their localized display names for different
585 * locales. The first element in the list is an array of the names
586 * of the public rule sets. The first element in this array is
587 * the initial default ruleset. The remaining elements in the
588 * list are arrays of localizations of the names of the public
589 * rule sets. Each of these is one longer than the initial array,
590 * with the first String being the ULocale ID, and the remaining
591 * Strings being the localizations of the rule set names, in the
592 * same order as the initial array. Arrays are NULL-terminated.
593 * @param rules A description of the formatter's desired behavior.
594 * See the class documentation for a complete explanation of the description
595 * syntax.
596 * @param localizations the localization information.
597 * names in the description. These will be copied by the constructor.
598 * @param perror The parse error if an error was encountered.
599 * @param status The status indicating whether the constructor succeeded.
600 * @stable ICU 3.2
601 */
602 RuleBasedNumberFormat(const UnicodeString& rules, const UnicodeString& localizations,
603 UParseError& perror, UErrorCode& status);
604
605 /**
606 * Creates a RuleBasedNumberFormat that behaves according to the rules
607 * passed in. The formatter uses the specified locale to determine the
608 * characters to use when formatting numerals, and to define equivalences
609 * for lenient parsing.
610 * @param rules The formatter rules.
611 * See the class documentation for a complete explanation of the rule
612 * syntax.
613 * @param locale A locale that governs which characters are used for
614 * formatting values in numerals and which characters are equivalent in
615 * lenient parsing.
616 * @param perror The parse error if an error was encountered.
617 * @param status The status indicating whether the constructor succeeded.
618 * @stable ICU 2.0
619 */
620 RuleBasedNumberFormat(const UnicodeString& rules, const Locale& locale,
621 UParseError& perror, UErrorCode& status);
622
623 /**
624 * Creates a RuleBasedNumberFormat that behaves according to the description
625 * passed in. The formatter uses the default locale.
626 * <p>
627 * The localizations data provides information about the public
628 * rule sets and their localized display names for different
629 * locales. The first element in the list is an array of the names
630 * of the public rule sets. The first element in this array is
631 * the initial default ruleset. The remaining elements in the
632 * list are arrays of localizations of the names of the public
633 * rule sets. Each of these is one longer than the initial array,
634 * with the first String being the ULocale ID, and the remaining
635 * Strings being the localizations of the rule set names, in the
636 * same order as the initial array. Arrays are NULL-terminated.
637 * @param rules A description of the formatter's desired behavior.
638 * See the class documentation for a complete explanation of the description
639 * syntax.
640 * @param localizations a list of localizations for the rule set
641 * names in the description. These will be copied by the constructor.
642 * @param locale A locale that governs which characters are used for
643 * formatting values in numerals and which characters are equivalent in
644 * lenient parsing.
645 * @param perror The parse error if an error was encountered.
646 * @param status The status indicating whether the constructor succeeded.
647 * @stable ICU 3.2
648 */
649 RuleBasedNumberFormat(const UnicodeString& rules, const UnicodeString& localizations,
650 const Locale& locale, UParseError& perror, UErrorCode& status);
651
652 /**
653 * Creates a RuleBasedNumberFormat from a predefined ruleset. The selector
654 * code choosed among three possible predefined formats: spellout, ordinal,
655 * and duration.
656 * @param tag A selector code specifying which kind of formatter to create for that
657 * locale. There are four legal values: URBNF_SPELLOUT, which creates a formatter that
658 * spells out a value in words in the desired language, URBNF_ORDINAL, which attaches
659 * an ordinal suffix from the desired language to the end of a number (e.g. "123rd"),
660 * URBNF_DURATION, which formats a duration in seconds as hours, minutes, and seconds always rounding down,
661 * and URBNF_NUMBERING_SYSTEM, which is used to invoke rules for alternate numbering
662 * systems such as the Hebrew numbering system, or for Roman Numerals, etc.
663 * @param locale The locale for the formatter.
664 * @param status The status indicating whether the constructor succeeded.
665 * @stable ICU 2.0
666 */
667 RuleBasedNumberFormat(URBNFRuleSetTag tag, const Locale& locale, UErrorCode& status);
668
669 //-----------------------------------------------------------------------
670 // boilerplate
671 //-----------------------------------------------------------------------
672
673 /**
674 * Copy constructor
675 * @param rhs the object to be copied from.
676 * @stable ICU 2.6
677 */
678 RuleBasedNumberFormat(const RuleBasedNumberFormat& rhs);
679
680 /**
681 * Assignment operator
682 * @param rhs the object to be copied from.
683 * @stable ICU 2.6
684 */
685 RuleBasedNumberFormat& operator=(const RuleBasedNumberFormat& rhs);
686
687 /**
688 * Release memory allocated for a RuleBasedNumberFormat when you are finished with it.
689 * @stable ICU 2.6
690 */
691 virtual ~RuleBasedNumberFormat();
692
693 /**
694 * Clone this object polymorphically. The caller is responsible
695 * for deleting the result when done.
696 * @return A copy of the object.
697 * @stable ICU 2.6
698 */
699 virtual Format* clone(void) const;
700
701 /**
702 * Return true if the given Format objects are semantically equal.
703 * Objects of different subclasses are considered unequal.
704 * @param other the object to be compared with.
705 * @return true if the given Format objects are semantically equal.
706 * @stable ICU 2.6
707 */
708 virtual UBool operator==(const Format& other) const;
709
710 //-----------------------------------------------------------------------
711 // public API functions
712 //-----------------------------------------------------------------------
713
714 /**
715 * return the rules that were provided to the RuleBasedNumberFormat.
716 * @return the result String that was passed in
717 * @stable ICU 2.0
718 */
719 virtual UnicodeString getRules() const;
720
721 /**
722 * Return the number of public rule set names.
723 * @return the number of public rule set names.
724 * @stable ICU 2.0
725 */
726 virtual int32_t getNumberOfRuleSetNames() const;
727
728 /**
729 * Return the name of the index'th public ruleSet. If index is not valid,
730 * the function returns null.
731 * @param index the index of the ruleset
732 * @return the name of the index'th public ruleSet.
733 * @stable ICU 2.0
734 */
735 virtual UnicodeString getRuleSetName(int32_t index) const;
736
737 /**
738 * Return the number of locales for which we have localized rule set display names.
739 * @return the number of locales for which we have localized rule set display names.
740 * @stable ICU 3.2
741 */
742 virtual int32_t getNumberOfRuleSetDisplayNameLocales(void) const;
743
744 /**
745 * Return the index'th display name locale.
746 * @param index the index of the locale
747 * @param status set to a failure code when this function fails
748 * @return the locale
749 * @see #getNumberOfRuleSetDisplayNameLocales
750 * @stable ICU 3.2
751 */
752 virtual Locale getRuleSetDisplayNameLocale(int32_t index, UErrorCode& status) const;
753
754 /**
755 * Return the rule set display names for the provided locale. These are in the same order
756 * as those returned by getRuleSetName. The locale is matched against the locales for
757 * which there is display name data, using normal fallback rules. If no locale matches,
758 * the default display names are returned. (These are the internal rule set names minus
759 * the leading '%'.)
760 * @param index the index of the rule set
761 * @param locale the locale (returned by getRuleSetDisplayNameLocales) for which the localized
762 * display name is desired
763 * @return the display name for the given index, which might be bogus if there is an error
764 * @see #getRuleSetName
765 * @stable ICU 3.2
766 */
767 virtual UnicodeString getRuleSetDisplayName(int32_t index,
768 const Locale& locale = Locale::getDefault());
769
770 /**
771 * Return the rule set display name for the provided rule set and locale.
772 * The locale is matched against the locales for which there is display name data, using
773 * normal fallback rules. If no locale matches, the default display name is returned.
774 * @return the display name for the rule set
775 * @stable ICU 3.2
776 * @see #getRuleSetDisplayName
777 */
778 virtual UnicodeString getRuleSetDisplayName(const UnicodeString& ruleSetName,
779 const Locale& locale = Locale::getDefault());
780
781
782 using NumberFormat::format;
783
784 /**
785 * Formats the specified 32-bit number using the default ruleset.
786 * @param number The number to format.
787 * @param toAppendTo the string that will hold the (appended) result
788 * @param pos the fieldposition
789 * @return A textual representation of the number.
790 * @stable ICU 2.0
791 */
792 virtual UnicodeString& format(int32_t number,
793 UnicodeString& toAppendTo,
794 FieldPosition& pos) const;
795
796 /**
797 * Formats the specified 64-bit number using the default ruleset.
798 * @param number The number to format.
799 * @param toAppendTo the string that will hold the (appended) result
800 * @param pos the fieldposition
801 * @return A textual representation of the number.
802 * @stable ICU 2.1
803 */
804 virtual UnicodeString& format(int64_t number,
805 UnicodeString& toAppendTo,
806 FieldPosition& pos) const;
807 /**
808 * Formats the specified number using the default ruleset.
809 * @param number The number to format.
810 * @param toAppendTo the string that will hold the (appended) result
811 * @param pos the fieldposition
812 * @return A textual representation of the number.
813 * @stable ICU 2.0
814 */
815 virtual UnicodeString& format(double number,
816 UnicodeString& toAppendTo,
817 FieldPosition& pos) const;
818
819 /**
820 * Formats the specified number using the named ruleset.
821 * @param number The number to format.
822 * @param ruleSetName The name of the rule set to format the number with.
823 * This must be the name of a valid public rule set for this formatter.
824 * @param toAppendTo the string that will hold the (appended) result
825 * @param pos the fieldposition
826 * @param status the status
827 * @return A textual representation of the number.
828 * @stable ICU 2.0
829 */
830 virtual UnicodeString& format(int32_t number,
831 const UnicodeString& ruleSetName,
832 UnicodeString& toAppendTo,
833 FieldPosition& pos,
834 UErrorCode& status) const;
835 /**
836 * Formats the specified 64-bit number using the named ruleset.
837 * @param number The number to format.
838 * @param ruleSetName The name of the rule set to format the number with.
839 * This must be the name of a valid public rule set for this formatter.
840 * @param toAppendTo the string that will hold the (appended) result
841 * @param pos the fieldposition
842 * @param status the status
843 * @return A textual representation of the number.
844 * @stable ICU 2.1
845 */
846 virtual UnicodeString& format(int64_t number,
847 const UnicodeString& ruleSetName,
848 UnicodeString& toAppendTo,
849 FieldPosition& pos,
850 UErrorCode& status) const;
851 /**
852 * Formats the specified number using the named ruleset.
853 * @param number The number to format.
854 * @param ruleSetName The name of the rule set to format the number with.
855 * This must be the name of a valid public rule set for this formatter.
856 * @param toAppendTo the string that will hold the (appended) result
857 * @param pos the fieldposition
858 * @param status the status
859 * @return A textual representation of the number.
860 * @stable ICU 2.0
861 */
862 virtual UnicodeString& format(double number,
863 const UnicodeString& ruleSetName,
864 UnicodeString& toAppendTo,
865 FieldPosition& pos,
866 UErrorCode& status) const;
867
868 protected:
869 /**
870 * Format a decimal number.
871 * The number is a DigitList wrapper onto a floating point decimal number.
872 * The default implementation in NumberFormat converts the decimal number
873 * to a double and formats that. Subclasses of NumberFormat that want
874 * to specifically handle big decimal numbers must override this method.
875 * class DecimalFormat does so.
876 *
877 * @param number The number, a DigitList format Decimal Floating Point.
878 * @param appendTo Output parameter to receive result.
879 * Result is appended to existing contents.
880 * @param posIter On return, can be used to iterate over positions
881 * of fields generated by this format call.
882 * @param status Output param filled with success/failure status.
883 * @return Reference to 'appendTo' parameter.
884 * @internal
885 */
886 virtual UnicodeString& format(const number::impl::DecimalQuantity &number,
887 UnicodeString& appendTo,
888 FieldPositionIterator* posIter,
889 UErrorCode& status) const;
890
891 /**
892 * Format a decimal number.
893 * The number is a DigitList wrapper onto a floating point decimal number.
894 * The default implementation in NumberFormat converts the decimal number
895 * to a double and formats that. Subclasses of NumberFormat that want
896 * to specifically handle big decimal numbers must override this method.
897 * class DecimalFormat does so.
898 *
899 * @param number The number, a DigitList format Decimal Floating Point.
900 * @param appendTo Output parameter to receive result.
901 * Result is appended to existing contents.
902 * @param pos On input: an alignment field, if desired.
903 * On output: the offsets of the alignment field.
904 * @param status Output param filled with success/failure status.
905 * @return Reference to 'appendTo' parameter.
906 * @internal
907 */
908 virtual UnicodeString& format(const number::impl::DecimalQuantity &number,
909 UnicodeString& appendTo,
910 FieldPosition& pos,
911 UErrorCode& status) const;
912 public:
913
914 using NumberFormat::parse;
915
916 /**
917 * Parses the specfied string, beginning at the specified position, according
918 * to this formatter's rules. This will match the string against all of the
919 * formatter's public rule sets and return the value corresponding to the longest
920 * parseable substring. This function's behavior is affected by the lenient
921 * parse mode.
922 * @param text The string to parse
923 * @param result the result of the parse, either a double or a long.
924 * @param parsePosition On entry, contains the position of the first character
925 * in "text" to examine. On exit, has been updated to contain the position
926 * of the first character in "text" that wasn't consumed by the parse.
927 * @see #setLenient
928 * @stable ICU 2.0
929 */
930 virtual void parse(const UnicodeString& text,
931 Formattable& result,
932 ParsePosition& parsePosition) const;
933
934 #if !UCONFIG_NO_COLLATION
935
936 /**
937 * Turns lenient parse mode on and off.
938 *
939 * When in lenient parse mode, the formatter uses a Collator for parsing the text.
940 * Only primary differences are treated as significant. This means that case
941 * differences, accent differences, alternate spellings of the same letter
942 * (e.g., ae and a-umlaut in German), ignorable characters, etc. are ignored in
943 * matching the text. In many cases, numerals will be accepted in place of words
944 * or phrases as well.
945 *
946 * For example, all of the following will correctly parse as 255 in English in
947 * lenient-parse mode:
948 * <br>"two hundred fifty-five"
949 * <br>"two hundred fifty five"
950 * <br>"TWO HUNDRED FIFTY-FIVE"
951 * <br>"twohundredfiftyfive"
952 * <br>"2 hundred fifty-5"
953 *
954 * The Collator used is determined by the locale that was
955 * passed to this object on construction. The description passed to this object
956 * on construction may supply additional collation rules that are appended to the
957 * end of the default collator for the locale, enabling additional equivalences
958 * (such as adding more ignorable characters or permitting spelled-out version of
959 * symbols; see the demo program for examples).
960 *
961 * It's important to emphasize that even strict parsing is relatively lenient: it
962 * will accept some text that it won't produce as output. In English, for example,
963 * it will correctly parse "two hundred zero" and "fifteen hundred".
964 *
965 * @param enabled If true, turns lenient-parse mode on; if false, turns it off.
966 * @see RuleBasedCollator
967 * @stable ICU 2.0
968 */
969 virtual void setLenient(UBool enabled);
970
971 /**
972 * Returns true if lenient-parse mode is turned on. Lenient parsing is off
973 * by default.
974 * @return true if lenient-parse mode is turned on.
975 * @see #setLenient
976 * @stable ICU 2.0
977 */
978 virtual inline UBool isLenient(void) const;
979
980 #endif
981
982 /**
983 * Override the default rule set to use. If ruleSetName is null, reset
984 * to the initial default rule set. If the rule set is not a public rule set name,
985 * U_ILLEGAL_ARGUMENT_ERROR is returned in status.
986 * @param ruleSetName the name of the rule set, or null to reset the initial default.
987 * @param status set to failure code when a problem occurs.
988 * @stable ICU 2.6
989 */
990 virtual void setDefaultRuleSet(const UnicodeString& ruleSetName, UErrorCode& status);
991
992 /**
993 * Return the name of the current default rule set. If the current rule set is
994 * not public, returns a bogus (and empty) UnicodeString.
995 * @return the name of the current default rule set
996 * @stable ICU 3.0
997 */
998 virtual UnicodeString getDefaultRuleSetName() const;
999
1000 /**
1001 * Set a particular UDisplayContext value in the formatter, such as
1002 * UDISPCTX_CAPITALIZATION_FOR_STANDALONE. Note: For getContext, see
1003 * NumberFormat.
1004 * @param value The UDisplayContext value to set.
1005 * @param status Input/output status. If at entry this indicates a failure
1006 * status, the function will do nothing; otherwise this will be
1007 * updated with any new status from the function.
1008 * @stable ICU 53
1009 */
1010 virtual void setContext(UDisplayContext value, UErrorCode& status);
1011
1012 /**
1013 * Get the rounding mode.
1014 * @return A rounding mode
1015 * @stable ICU 60
1016 */
1017 virtual ERoundingMode getRoundingMode(void) const;
1018
1019 /**
1020 * Set the rounding mode.
1021 * @param roundingMode A rounding mode
1022 * @stable ICU 60
1023 */
1024 virtual void setRoundingMode(ERoundingMode roundingMode);
1025
1026 public:
1027 /**
1028 * ICU "poor man's RTTI", returns a UClassID for this class.
1029 *
1030 * @stable ICU 2.8
1031 */
1032 static UClassID U_EXPORT2 getStaticClassID(void);
1033
1034 /**
1035 * ICU "poor man's RTTI", returns a UClassID for the actual class.
1036 *
1037 * @stable ICU 2.8
1038 */
1039 virtual UClassID getDynamicClassID(void) const;
1040
1041 /**
1042 * Sets the decimal format symbols, which is generally not changed
1043 * by the programmer or user. The formatter takes ownership of
1044 * symbolsToAdopt; the client must not delete it.
1045 *
1046 * @param symbolsToAdopt DecimalFormatSymbols to be adopted.
1047 * @stable ICU 49
1048 */
1049 virtual void adoptDecimalFormatSymbols(DecimalFormatSymbols* symbolsToAdopt);
1050
1051 /**
1052 * Sets the decimal format symbols, which is generally not changed
1053 * by the programmer or user. A clone of the symbols is created and
1054 * the symbols is _not_ adopted; the client is still responsible for
1055 * deleting it.
1056 *
1057 * @param symbols DecimalFormatSymbols.
1058 * @stable ICU 49
1059 */
1060 virtual void setDecimalFormatSymbols(const DecimalFormatSymbols& symbols);
1061
1062 private:
1063 RuleBasedNumberFormat(); // default constructor not implemented
1064
1065 // this will ref the localizations if they are not NULL
1066 // caller must deref to get adoption
1067 RuleBasedNumberFormat(const UnicodeString& description, LocalizationInfo* localizations,
1068 const Locale& locale, UParseError& perror, UErrorCode& status);
1069
1070 void init(const UnicodeString& rules, LocalizationInfo* localizations, UParseError& perror, UErrorCode& status);
1071 void initCapitalizationContextInfo(const Locale& thelocale);
1072 void dispose();
1073 void stripWhitespace(UnicodeString& src);
1074 void initDefaultRuleSet();
1075 NFRuleSet* findRuleSet(const UnicodeString& name, UErrorCode& status) const;
1076
1077 /* friend access */
1078 friend class NFSubstitution;
1079 friend class NFRule;
1080 friend class NFRuleSet;
1081 friend class FractionalPartSubstitution;
1082
1083 inline NFRuleSet * getDefaultRuleSet() const;
1084 const RuleBasedCollator * getCollator() const;
1085 DecimalFormatSymbols * initializeDecimalFormatSymbols(UErrorCode &status);
1086 const DecimalFormatSymbols * getDecimalFormatSymbols() const;
1087 NFRule * initializeDefaultInfinityRule(UErrorCode &status);
1088 const NFRule * getDefaultInfinityRule() const;
1089 NFRule * initializeDefaultNaNRule(UErrorCode &status);
1090 const NFRule * getDefaultNaNRule() const;
1091 PluralFormat *createPluralFormat(UPluralType pluralType, const UnicodeString &pattern, UErrorCode& status) const;
1092 UnicodeString& adjustForCapitalizationContext(int32_t startPos, UnicodeString& currentResult, UErrorCode& status) const;
1093 UnicodeString& format(int64_t number, NFRuleSet *ruleSet, UnicodeString& toAppendTo, UErrorCode& status) const;
1094 void format(double number, NFRuleSet& rs, UnicodeString& toAppendTo, UErrorCode& status) const;
1095
1096 private:
1097 NFRuleSet **fRuleSets;
1098 UnicodeString* ruleSetDescriptions;
1099 int32_t numRuleSets;
1100 NFRuleSet *defaultRuleSet;
1101 Locale locale;
1102 RuleBasedCollator* collator;
1103 DecimalFormatSymbols* decimalFormatSymbols;
1104 NFRule *defaultInfinityRule;
1105 NFRule *defaultNaNRule;
1106 ERoundingMode fRoundingMode;
1107 UBool lenient;
1108 UnicodeString* lenientParseRules;
1109 LocalizationInfo* localizations;
1110 UnicodeString originalDescription;
1111 UBool capitalizationInfoSet;
1112 UBool capitalizationForUIListMenu;
1113 UBool capitalizationForStandAlone;
1114 BreakIterator* capitalizationBrkIter;
1115 };
1116
1117 // ---------------
1118
1119 #if !UCONFIG_NO_COLLATION
1120
1121 inline UBool
isLenient(void)1122 RuleBasedNumberFormat::isLenient(void) const {
1123 return lenient;
1124 }
1125
1126 #endif
1127
1128 inline NFRuleSet*
getDefaultRuleSet()1129 RuleBasedNumberFormat::getDefaultRuleSet() const {
1130 return defaultRuleSet;
1131 }
1132
1133 U_NAMESPACE_END
1134
1135 /* U_HAVE_RBNF */
1136 #endif
1137
1138 /* RBNF_H */
1139 #endif
1140