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 #if U_SHOW_CPLUSPLUS_API
16
17 /**
18 * \file
19 * \brief C++ API: Rule Based Number Format
20 */
21
22 /**
23 * \def U_HAVE_RBNF
24 * This will be 0 if RBNF support is not included in ICU
25 * and 1 if it is.
26 *
27 * @stable ICU 2.4
28 */
29 #if UCONFIG_NO_FORMATTING
30 #define U_HAVE_RBNF 0
31 #else
32 #define U_HAVE_RBNF 1
33
34 #include "unicode/dcfmtsym.h"
35 #include "unicode/fmtable.h"
36 #include "unicode/locid.h"
37 #include "unicode/numfmt.h"
38 #include "unicode/unistr.h"
39 #include "unicode/strenum.h"
40 #include "unicode/brkiter.h"
41 #include "unicode/upluralrules.h"
42
43 U_NAMESPACE_BEGIN
44
45 class NFRule;
46 class NFRuleSet;
47 class LocalizationInfo;
48 class PluralFormat;
49 class RuleBasedCollator;
50
51 /**
52 * Tags for the predefined rulesets.
53 *
54 * @stable ICU 2.2
55 */
56 enum URBNFRuleSetTag {
57 /**
58 * Requests predefined ruleset for spelling out numeric values in words.
59 * @stable ICU 2.2
60 */
61 URBNF_SPELLOUT,
62 /**
63 * Requests predefined ruleset for the ordinal form of a number.
64 * @stable ICU 2.2
65 */
66 URBNF_ORDINAL,
67 #ifndef U_HIDE_DEPRECATED_API
68 /**
69 * Requests predefined ruleset for formatting a value as a duration in hours, minutes, and seconds.
70 * @deprecated ICU 74 Use MeasureFormat instead.
71 */
72 URBNF_DURATION,
73 #endif // U_HIDE_DERECATED_API
74 /**
75 * Requests predefined ruleset for various non-place-value numbering systems.
76 * WARNING: The same resource contains rule sets for a variety of different numbering systems.
77 * You need to call setDefaultRuleSet() on the formatter to choose the actual numbering system.
78 * @stable ICU 2.2
79 */
80 URBNF_NUMBERING_SYSTEM = 3,
81 #ifndef U_HIDE_DEPRECATED_API
82 /**
83 * One more than the highest normal URBNFRuleSetTag value.
84 * @deprecated ICU 58 The numeric value may change over time, see ICU ticket #12420.
85 */
86 URBNF_COUNT
87 #endif // U_HIDE_DEPRECATED_API
88 };
89
90 /**
91 * The RuleBasedNumberFormat class formats numbers according to a set of rules. This number formatter is
92 * typically used for spelling out numeric values in words (e.g., 25,3476 as
93 * "twenty-five thousand three hundred seventy-six" or "vingt-cinq mille trois
94 * cents soixante-seize" or
95 * "fünfundzwanzigtausenddreihundertsechsundsiebzig"), but can also be used for
96 * other complicated formatting tasks, such as formatting a number of seconds as hours,
97 * minutes and seconds (e.g., 3,730 as "1:02:10").
98 *
99 * <p>The resources contain three predefined formatters for each locale: spellout, which
100 * spells out a value in words (123 is "one hundred twenty-three"); ordinal, which
101 * appends an ordinal suffix to the end of a numeral (123 is "123rd"); and
102 * duration, which shows a duration in seconds as hours, minutes, and seconds (123 is
103 * "2:03"). The client can also define more specialized <tt>RuleBasedNumberFormat</tt>s
104 * by supplying programmer-defined rule sets.</p>
105 *
106 * <p>The behavior of a <tt>RuleBasedNumberFormat</tt> is specified by a textual description
107 * that is either passed to the constructor as a <tt>String</tt> or loaded from a resource
108 * bundle. In its simplest form, the description consists of a semicolon-delimited list of <em>rules.</em>
109 * Each rule has a string of output text and a value or range of values it is applicable to.
110 * In a typical spellout rule set, the first twenty rules are the words for the numbers from
111 * 0 to 19:</p>
112 *
113 * <pre>zero; one; two; three; four; five; six; seven; eight; nine;
114 * ten; eleven; twelve; thirteen; fourteen; fifteen; sixteen; seventeen; eighteen; nineteen;</pre>
115 *
116 * <p>For larger numbers, we can use the preceding set of rules to format the ones place, and
117 * we only have to supply the words for the multiples of 10:</p>
118 *
119 * <pre> 20: twenty[->>];
120 * 30: thirty[->>];
121 * 40: forty[->>];
122 * 50: fifty[->>];
123 * 60: sixty[->>];
124 * 70: seventy[->>];
125 * 80: eighty[->>];
126 * 90: ninety[->>];</pre>
127 *
128 * <p>In these rules, the <em>base value</em> is spelled out explicitly and set off from the
129 * rule's output text with a colon. The rules are in a sorted list, and a rule is applicable
130 * to all numbers from its own base value to one less than the next rule's base value. The
131 * ">>" token is called a <em>substitution</em> and tells the formatter to
132 * isolate the number's ones digit, format it using this same set of rules, and place the
133 * result at the position of the ">>" token. Text in brackets is omitted if
134 * the number being formatted is an even multiple of 10 (the hyphen is a literal hyphen; 24
135 * is "twenty-four," not "twenty four").</p>
136 *
137 * <p>For even larger numbers, we can actually look up several parts of the number in the
138 * list:</p>
139 *
140 * <pre>100: << hundred[ >>];</pre>
141 *
142 * <p>The "<<" represents a new kind of substitution. The << isolates
143 * the hundreds digit (and any digits to its left), formats it using this same rule set, and
144 * places the result where the "<<" was. Notice also that the meaning of
145 * >> has changed: it now refers to both the tens and the ones digits. The meaning of
146 * both substitutions depends on the rule's base value. The base value determines the rule's <em>divisor,</em>
147 * which is the highest power of 10 that is less than or equal to the base value (the user
148 * can change this). To fill in the substitutions, the formatter divides the number being
149 * formatted by the divisor. The integral quotient is used to fill in the <<
150 * substitution, and the remainder is used to fill in the >> substitution. The meaning
151 * of the brackets changes similarly: text in brackets is omitted if the value being
152 * formatted is an even multiple of the rule's divisor. The rules are applied recursively, so
153 * if a substitution is filled in with text that includes another substitution, that
154 * substitution is also filled in.</p>
155 *
156 * <p>This rule covers values up to 999, at which point we add another rule:</p>
157 *
158 * <pre>1000: << thousand[ >>];</pre>
159 *
160 * <p>Again, the meanings of the brackets and substitution tokens shift because the rule's
161 * base value is a higher power of 10, changing the rule's divisor. This rule can actually be
162 * used all the way up to 999,999. This allows us to finish out the rules as follows:</p>
163 *
164 * <pre> 1,000,000: << million[ >>];
165 * 1,000,000,000: << billion[ >>];
166 * 1,000,000,000,000: << trillion[ >>];
167 * 1,000,000,000,000,000: OUT OF RANGE!;</pre>
168 *
169 * <p>Commas, periods, and spaces can be used in the base values to improve legibility and
170 * are ignored by the rule parser. The last rule in the list is customarily treated as an
171 * "overflow rule," applying to everything from its base value on up, and often (as
172 * in this example) being used to print out an error message or default representation.
173 * Notice also that the size of the major groupings in large numbers is controlled by the
174 * spacing of the rules: because in English we group numbers by thousand, the higher rules
175 * are separated from each other by a factor of 1,000.</p>
176 *
177 * <p>To see how these rules actually work in practice, consider the following example:
178 * Formatting 25,430 with this rule set would work like this:</p>
179 *
180 * <table border="0" width="100%">
181 * <tr>
182 * <td><strong><< thousand >></strong></td>
183 * <td>[the rule whose base value is 1,000 is applicable to 25,340]</td>
184 * </tr>
185 * <tr>
186 * <td><strong>twenty->></strong> thousand >></td>
187 * <td>[25,340 over 1,000 is 25. The rule for 20 applies.]</td>
188 * </tr>
189 * <tr>
190 * <td>twenty-<strong>five</strong> thousand >></td>
191 * <td>[25 mod 10 is 5. The rule for 5 is "five."</td>
192 * </tr>
193 * <tr>
194 * <td>twenty-five thousand <strong><< hundred >></strong></td>
195 * <td>[25,340 mod 1,000 is 340. The rule for 100 applies.]</td>
196 * </tr>
197 * <tr>
198 * <td>twenty-five thousand <strong>three</strong> hundred >></td>
199 * <td>[340 over 100 is 3. The rule for 3 is "three."]</td>
200 * </tr>
201 * <tr>
202 * <td>twenty-five thousand three hundred <strong>forty</strong></td>
203 * <td>[340 mod 100 is 40. The rule for 40 applies. Since 40 divides
204 * evenly by 10, the hyphen and substitution in the brackets are omitted.]</td>
205 * </tr>
206 * </table>
207 *
208 * <p>The above syntax suffices only to format positive integers. To format negative numbers,
209 * we add a special rule:</p>
210 *
211 * <pre>-x: minus >>;</pre>
212 *
213 * <p>This is called a <em>negative-number rule,</em> and is identified by "-x"
214 * where the base value would be. This rule is used to format all negative numbers. the
215 * >> token here means "find the number's absolute value, format it with these
216 * rules, and put the result here."</p>
217 *
218 * <p>We also add a special rule called a <em>fraction rule </em>for numbers with fractional
219 * parts:</p>
220 *
221 * <pre>x.x: << point >>;</pre>
222 *
223 * <p>This rule is used for all positive non-integers (negative non-integers pass through the
224 * negative-number rule first and then through this rule). Here, the << token refers to
225 * the number's integral part, and the >> to the number's fractional part. The
226 * fractional part is formatted as a series of single-digit numbers (e.g., 123.456 would be
227 * formatted as "one hundred twenty-three point four five six").</p>
228 *
229 * <p>To see how this rule syntax is applied to various languages, examine the resource data.</p>
230 *
231 * <p>There is actually much more flexibility built into the rule language than the
232 * description above shows. A formatter may own multiple rule sets, which can be selected by
233 * the caller, and which can use each other to fill in their substitutions. Substitutions can
234 * also be filled in with digits, using a DecimalFormat object. There is syntax that can be
235 * used to alter a rule's divisor in various ways. And there is provision for much more
236 * flexible fraction handling. A complete description of the rule syntax follows:</p>
237 *
238 * <hr>
239 *
240 * <p>The description of a <tt>RuleBasedNumberFormat</tt>'s behavior consists of one or more <em>rule
241 * sets.</em> Each rule set consists of a name, a colon, and a list of <em>rules.</em> A rule
242 * set name must begin with a % sign. Rule sets with names that begin with a single % sign
243 * are <em>public:</em> the caller can specify that they be used to format and parse numbers.
244 * Rule sets with names that begin with %% are <em>private:</em> they exist only for the use
245 * of other rule sets. If a formatter only has one rule set, the name may be omitted.</p>
246 *
247 * <p>The user can also specify a special "rule set" named <tt>%%lenient-parse</tt>.
248 * The body of <tt>%%lenient-parse</tt> isn't a set of number-formatting rules, but a <tt>RuleBasedCollator</tt>
249 * description which is used to define equivalences for lenient parsing. For more information
250 * on the syntax, see <tt>RuleBasedCollator</tt>. For more information on lenient parsing,
251 * see <tt>setLenientParse()</tt>. <em>Note:</em> symbols that have syntactic meaning
252 * in collation rules, such as '&', have no particular meaning when appearing outside
253 * of the <tt>lenient-parse</tt> rule set.</p>
254 *
255 * <p>The body of a rule set consists of an ordered, semicolon-delimited list of <em>rules.</em>
256 * Internally, every rule has a base value, a divisor, rule text, and zero, one, or two <em>substitutions.</em>
257 * These parameters are controlled by the description syntax, which consists of a <em>rule
258 * descriptor,</em> a colon, and a <em>rule body.</em></p>
259 *
260 * <p>A rule descriptor can take one of the following forms (text in <em>italics</em> is the
261 * name of a token):</p>
262 *
263 * <table border="0" width="100%">
264 * <tr>
265 * <td><em>bv</em>:</td>
266 * <td><em>bv</em> specifies the rule's base value. <em>bv</em> is a decimal
267 * number expressed using ASCII digits. <em>bv</em> may contain spaces, period, and commas,
268 * which are ignored. The rule's divisor is the highest power of 10 less than or equal to
269 * the base value.</td>
270 * </tr>
271 * <tr>
272 * <td><em>bv</em>/<em>rad</em>:</td>
273 * <td><em>bv</em> specifies the rule's base value. The rule's divisor is the
274 * highest power of <em>rad</em> less than or equal to the base value.</td>
275 * </tr>
276 * <tr>
277 * <td><em>bv</em>>:</td>
278 * <td><em>bv</em> specifies the rule's base value. To calculate the divisor,
279 * let the radix be 10, and the exponent be the highest exponent of the radix that yields a
280 * result less than or equal to the base value. Every > character after the base value
281 * decreases the exponent by 1. If the exponent is positive or 0, the divisor is the radix
282 * raised to the power of the exponent; otherwise, the divisor is 1.</td>
283 * </tr>
284 * <tr>
285 * <td><em>bv</em>/<em>rad</em>>:</td>
286 * <td><em>bv</em> specifies the rule's base value. To calculate the divisor,
287 * let the radix be <em>rad</em>, and the exponent be the highest exponent of the radix that
288 * yields a result less than or equal to the base value. Every > character after the radix
289 * decreases the exponent by 1. If the exponent is positive or 0, the divisor is the radix
290 * raised to the power of the exponent; otherwise, the divisor is 1.</td>
291 * </tr>
292 * <tr>
293 * <td>-x:</td>
294 * <td>The rule is a negative-number rule.</td>
295 * </tr>
296 * <tr>
297 * <td>x.x:</td>
298 * <td>The rule is an <em>improper fraction 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,x instead. For example,
303 * you can use "x.x: << point >>;x,x: << 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>0.x:</td>
309 * <td>The rule is a <em>proper fraction rule</em>. If the full stop in
310 * the middle of the rule name is replaced with the decimal point
311 * that is used in the language or DecimalFormatSymbols, then that rule will
312 * have precedence when formatting and parsing this rule. For example, some
313 * languages use the comma, and can thus be written as 0,x instead. For example,
314 * you can use "0.x: point >>;0,x: comma >>;" to
315 * handle the decimal point that matches the language's natural spelling of
316 * the punctuation of either the full stop or comma.</td>
317 * </tr>
318 * <tr>
319 * <td>x.0:</td>
320 * <td>The rule is a <em>default rule</em>. If the full stop in
321 * the middle of the rule name is replaced with the decimal point
322 * that is used in the language or DecimalFormatSymbols, then that rule will
323 * have precedence when formatting and parsing this rule. For example, some
324 * languages use the comma, and can thus be written as x,0 instead. For example,
325 * you can use "x.0: << point;x,0: << comma;" to
326 * handle the decimal point that matches the language's natural spelling of
327 * the punctuation of either the full stop or comma.</td>
328 * </tr>
329 * <tr>
330 * <td>Inf:</td>
331 * <td>The rule for infinity.</td>
332 * </tr>
333 * <tr>
334 * <td>NaN:</td>
335 * <td>The rule for an IEEE 754 NaN (not a number).</td>
336 * </tr>
337 * <tr>
338 * <td><em>nothing</em></td>
339 * <td>If the rule's rule descriptor is left out, the base value is one plus the
340 * preceding rule's base value (or zero if this is the first rule in the list) in a normal
341 * rule set. In a fraction rule set, the base value is the same as the preceding rule's
342 * base value.</td>
343 * </tr>
344 * </table>
345 *
346 * <p>A rule set may be either a regular rule set or a <em>fraction rule set,</em> depending
347 * on whether it is used to format a number's integral part (or the whole number) or a
348 * number's fractional part. Using a rule set to format a rule's fractional part makes it a
349 * fraction rule set.</p>
350 *
351 * <p>Which rule is used to format a number is defined according to one of the following
352 * algorithms: If the rule set is a regular rule set, do the following:
353 *
354 * <ul>
355 * <li>If the rule set includes a default rule (and the number was passed in as a <tt>double</tt>),
356 * use the default rule. (If the number being formatted was passed in as a <tt>long</tt>,
357 * the default rule is ignored.)</li>
358 * <li>If the number is negative, use the negative-number rule.</li>
359 * <li>If the number has a fractional part and is greater than 1, use the improper fraction
360 * rule.</li>
361 * <li>If the number has a fractional part and is between 0 and 1, use the proper fraction
362 * rule.</li>
363 * <li>Binary-search the rule list for the rule with the highest base value less than or equal
364 * to the number. If that rule has two substitutions, its base value is not an even multiple
365 * of its divisor, and the number <em>is</em> an even multiple of the rule's divisor, use the
366 * rule that precedes it in the rule list. Otherwise, use the rule itself.</li>
367 * </ul>
368 *
369 * <p>If the rule set is a fraction rule set, do the following:
370 *
371 * <ul>
372 * <li>Ignore negative-number and fraction rules.</li>
373 * <li>For each rule in the list, multiply the number being formatted (which will always be
374 * between 0 and 1) by the rule's base value. Keep track of the distance between the result
375 * the nearest integer.</li>
376 * <li>Use the rule that produced the result closest to zero in the above calculation. In the
377 * event of a tie or a direct hit, use the first matching rule encountered. (The idea here is
378 * to try each rule's base value as a possible denominator of a fraction. Whichever
379 * denominator produces the fraction closest in value to the number being formatted wins.) If
380 * the rule following the matching rule has the same base value, use it if the numerator of
381 * the fraction is anything other than 1; if the numerator is 1, use the original matching
382 * rule. (This is to allow singular and plural forms of the rule text without a lot of extra
383 * hassle.)</li>
384 * </ul>
385 *
386 * <p>A rule's body consists of a string of characters terminated by a semicolon. The rule
387 * may include zero, one, or two <em>substitution tokens,</em> and a range of text in
388 * brackets. The brackets denote optional text (and may also include one or both
389 * substitutions). The exact meanings of the substitution tokens, and under what conditions
390 * optional text is omitted, depend on the syntax of the substitution token and the context.
391 * The rest of the text in a rule body is literal text that is output when the rule matches
392 * the number being formatted.</p>
393 *
394 * <p>A substitution token begins and ends with a <em>token character.</em> The token
395 * character and the context together specify a mathematical operation to be performed on the
396 * number being formatted. An optional <em>substitution descriptor </em>specifies how the
397 * value resulting from that operation is used to fill in the substitution. The position of
398 * the substitution token in the rule body specifies the location of the resultant text in
399 * the original rule text.</p>
400 *
401 * <p>The meanings of the substitution token characters are as follows:</p>
402 *
403 * <table border="0" width="100%">
404 * <tr>
405 * <td>>></td>
406 * <td>in normal rule</td>
407 * <td>Divide the number by the rule's divisor and format the remainder</td>
408 * </tr>
409 * <tr>
410 * <td></td>
411 * <td>in negative-number rule</td>
412 * <td>Find the absolute value of the number and format the result</td>
413 * </tr>
414 * <tr>
415 * <td></td>
416 * <td>in fraction or default rule</td>
417 * <td>Isolate the number's fractional part and format it.</td>
418 * </tr>
419 * <tr>
420 * <td></td>
421 * <td>in rule in fraction rule set</td>
422 * <td>Not allowed.</td>
423 * </tr>
424 * <tr>
425 * <td>>>></td>
426 * <td>in normal rule</td>
427 * <td>Divide the number by the rule's divisor and format the remainder,
428 * but bypass the normal rule-selection process and just use the
429 * rule that precedes this one in this rule list.</td>
430 * </tr>
431 * <tr>
432 * <td></td>
433 * <td>in all other rules</td>
434 * <td>Not allowed.</td>
435 * </tr>
436 * <tr>
437 * <td><<</td>
438 * <td>in normal rule</td>
439 * <td>Divide the number by the rule's divisor and format the quotient</td>
440 * </tr>
441 * <tr>
442 * <td></td>
443 * <td>in negative-number rule</td>
444 * <td>Not allowed.</td>
445 * </tr>
446 * <tr>
447 * <td></td>
448 * <td>in fraction or default rule</td>
449 * <td>Isolate the number's integral part and format it.</td>
450 * </tr>
451 * <tr>
452 * <td></td>
453 * <td>in rule in fraction rule set</td>
454 * <td>Multiply the number by the rule's base value and format the result.</td>
455 * </tr>
456 * <tr>
457 * <td>==</td>
458 * <td>in all rule sets</td>
459 * <td>Format the number unchanged</td>
460 * </tr>
461 * <tr>
462 * <td>[]</td>
463 * <td>in normal rule</td>
464 * <td>Omit the optional text if the number is an even multiple of the rule's divisor</td>
465 * </tr>
466 * <tr>
467 * <td></td>
468 * <td>in negative-number rule</td>
469 * <td>Not allowed.</td>
470 * </tr>
471 * <tr>
472 * <td></td>
473 * <td>in improper-fraction rule</td>
474 * <td>Omit the optional text if the number is between 0 and 1 (same as specifying both an
475 * x.x rule and a 0.x rule)</td>
476 * </tr>
477 * <tr>
478 * <td></td>
479 * <td>in default rule</td>
480 * <td>Omit the optional text if the number is an integer (same as specifying both an x.x
481 * rule and an x.0 rule)</td>
482 * </tr>
483 * <tr>
484 * <td></td>
485 * <td>in proper-fraction rule</td>
486 * <td>Not allowed.</td>
487 * </tr>
488 * <tr>
489 * <td></td>
490 * <td>in rule in fraction rule set</td>
491 * <td>Omit the optional text if multiplying the number by the rule's base value yields 1.</td>
492 * </tr>
493 * <tr>
494 * <td width="37">$(cardinal,<i>plural syntax</i>)$</td>
495 * <td width="23"></td>
496 * <td width="165" valign="top">in all rule sets</td>
497 * <td>This provides the ability to choose a word based on the number divided by the radix to the power of the
498 * exponent of the base value for the specified locale, which is normally equivalent to the << value.
499 * This uses the cardinal plural rules from PluralFormat. All strings used in the plural format are treated
500 * as the same base value for parsing.</td>
501 * </tr>
502 * <tr>
503 * <td width="37">$(ordinal,<i>plural syntax</i>)$</td>
504 * <td width="23"></td>
505 * <td width="165" valign="top">in all rule sets</td>
506 * <td>This provides the ability to choose a word based on the number divided by the radix to the power of the
507 * exponent of the base value for the specified locale, which is normally equivalent to the << value.
508 * This uses the ordinal plural rules from PluralFormat. All strings used in the plural format are treated
509 * as the same base value for parsing.</td>
510 * </tr>
511 * </table>
512 *
513 * <p>The substitution descriptor (i.e., the text between the token characters) may take one
514 * of three forms:</p>
515 *
516 * <table border="0" width="100%">
517 * <tr>
518 * <td>a rule set name</td>
519 * <td>Perform the mathematical operation on the number, and format the result using the
520 * named rule set.</td>
521 * </tr>
522 * <tr>
523 * <td>a DecimalFormat pattern</td>
524 * <td>Perform the mathematical operation on the number, and format the result using a
525 * DecimalFormat with the specified pattern. The pattern must begin with 0 or #.</td>
526 * </tr>
527 * <tr>
528 * <td>nothing</td>
529 * <td>Perform the mathematical operation on the number, and format the result using the rule
530 * set containing the current rule, except:
531 * <ul>
532 * <li>You can't have an empty substitution descriptor with a == substitution.</li>
533 * <li>If you omit the substitution descriptor in a >> substitution in a fraction rule,
534 * format the result one digit at a time using the rule set containing the current rule.</li>
535 * <li>If you omit the substitution descriptor in a << substitution in a rule in a
536 * fraction rule set, format the result using the default rule set for this formatter.</li>
537 * </ul>
538 * </td>
539 * </tr>
540 * </table>
541 *
542 * <p>Whitespace is ignored between a rule set name and a rule set body, between a rule
543 * descriptor and a rule body, or between rules. If a rule body begins with an apostrophe,
544 * the apostrophe is ignored, but all text after it becomes significant (this is how you can
545 * have a rule's rule text begin with whitespace). There is no escape function: the semicolon
546 * is not allowed in rule set names or in rule text, and the colon is not allowed in rule set
547 * names. The characters beginning a substitution token are always treated as the beginning
548 * of a substitution token.</p>
549 *
550 * <p>See the resource data and the demo program for annotated examples of real rule sets
551 * using these features.</p>
552 *
553 * <p><em>User subclasses are not supported.</em> While clients may write
554 * subclasses, such code will not necessarily work and will not be
555 * guaranteed to work stably from release to release.
556 *
557 * <p><b>Localizations</b></p>
558 * <p>Constructors are available that allow the specification of localizations for the
559 * public rule sets (and also allow more control over what public rule sets are available).
560 * Localization data is represented as a textual description. The description represents
561 * an array of arrays of string. The first element is an array of the public rule set names,
562 * each of these must be one of the public rule set names that appear in the rules. Only
563 * names in this array will be treated as public rule set names by the API. Each subsequent
564 * element is an array of localizations of these names. The first element of one of these
565 * subarrays is the locale name, and the remaining elements are localizations of the
566 * public rule set names, in the same order as they were listed in the first array.</p>
567 * <p>In the syntax, angle brackets '<', '>' are used to delimit the arrays, and comma ',' is used
568 * to separate elements of an array. Whitespace is ignored, unless quoted.</p>
569 * <p>For example:<pre>
570 * < < %foo, %bar, %baz >,
571 * < en, Foo, Bar, Baz >,
572 * < fr, 'le Foo', 'le Bar', 'le Baz' >
573 * < zh, \\u7532, \\u4e59, \\u4e19 > >
574 * </pre></p>
575 * @author Richard Gillam
576 * @see NumberFormat
577 * @see DecimalFormat
578 * @see PluralFormat
579 * @see PluralRules
580 * @stable ICU 2.0
581 */
582 class U_I18N_API RuleBasedNumberFormat : public NumberFormat {
583 public:
584
585 //-----------------------------------------------------------------------
586 // constructors
587 //-----------------------------------------------------------------------
588
589 /**
590 * Creates a RuleBasedNumberFormat that behaves according to the description
591 * passed in. The formatter uses the default locale.
592 * @param rules A description of the formatter's desired behavior.
593 * See the class documentation for a complete explanation of the description
594 * syntax.
595 * @param perror The parse error if an error was encountered.
596 * @param status The status indicating whether the constructor succeeded.
597 * @stable ICU 3.2
598 */
599 RuleBasedNumberFormat(const UnicodeString& rules, UParseError& perror, UErrorCode& status);
600
601 /**
602 * Creates a RuleBasedNumberFormat that behaves according to the description
603 * passed in. The formatter uses the default locale.
604 * <p>
605 * The localizations data provides information about the public
606 * rule sets and their localized display names for different
607 * locales. The first element in the list is an array of the names
608 * of the public rule sets. The first element in this array is
609 * the initial default ruleset. The remaining elements in the
610 * list are arrays of localizations of the names of the public
611 * rule sets. Each of these is one longer than the initial array,
612 * with the first String being the ULocale ID, and the remaining
613 * Strings being the localizations of the rule set names, in the
614 * same order as the initial array. Arrays are nullptr-terminated.
615 * @param rules A description of the formatter's desired behavior.
616 * See the class documentation for a complete explanation of the description
617 * syntax.
618 * @param localizations the localization information.
619 * names in the description. These will be copied by the constructor.
620 * @param perror The parse error if an error was encountered.
621 * @param status The status indicating whether the constructor succeeded.
622 * @stable ICU 3.2
623 */
624 RuleBasedNumberFormat(const UnicodeString& rules, const UnicodeString& localizations,
625 UParseError& perror, UErrorCode& status);
626
627 /**
628 * Creates a RuleBasedNumberFormat that behaves according to the rules
629 * passed in. The formatter uses the specified locale to determine the
630 * characters to use when formatting numerals, and to define equivalences
631 * for lenient parsing.
632 * @param rules The formatter rules.
633 * See the class documentation for a complete explanation of the rule
634 * syntax.
635 * @param locale A locale that governs which characters are used for
636 * formatting values in numerals and which characters are equivalent in
637 * lenient parsing.
638 * @param perror The parse error if an error was encountered.
639 * @param status The status indicating whether the constructor succeeded.
640 * @stable ICU 2.0
641 */
642 RuleBasedNumberFormat(const UnicodeString& rules, const Locale& locale,
643 UParseError& perror, UErrorCode& status);
644
645 /**
646 * Creates a RuleBasedNumberFormat that behaves according to the description
647 * passed in. The formatter uses the default locale.
648 * <p>
649 * The localizations data provides information about the public
650 * rule sets and their localized display names for different
651 * locales. The first element in the list is an array of the names
652 * of the public rule sets. The first element in this array is
653 * the initial default ruleset. The remaining elements in the
654 * list are arrays of localizations of the names of the public
655 * rule sets. Each of these is one longer than the initial array,
656 * with the first String being the ULocale ID, and the remaining
657 * Strings being the localizations of the rule set names, in the
658 * same order as the initial array. Arrays are nullptr-terminated.
659 * @param rules A description of the formatter's desired behavior.
660 * See the class documentation for a complete explanation of the description
661 * syntax.
662 * @param localizations a list of localizations for the rule set
663 * names in the description. These will be copied by the constructor.
664 * @param locale A locale that governs which characters are used for
665 * formatting values in numerals and which characters are equivalent in
666 * lenient parsing.
667 * @param perror The parse error if an error was encountered.
668 * @param status The status indicating whether the constructor succeeded.
669 * @stable ICU 3.2
670 */
671 RuleBasedNumberFormat(const UnicodeString& rules, const UnicodeString& localizations,
672 const Locale& locale, UParseError& perror, UErrorCode& status);
673
674 /**
675 * Creates a RuleBasedNumberFormat from a predefined ruleset. The selector
676 * code chose among three possible predefined formats: spellout, ordinal,
677 * and duration.
678 * @param tag A selector code specifying which kind of formatter to create for that
679 * locale. There are four legal values: URBNF_SPELLOUT, which creates a formatter that
680 * spells out a value in words in the desired language, URBNF_ORDINAL, which attaches
681 * an ordinal suffix from the desired language to the end of a number (e.g. "123rd"),
682 * URBNF_DURATION, which formats a duration in seconds as hours, minutes, and seconds always rounding down,
683 * and URBNF_NUMBERING_SYSTEM, which is used to invoke rules for alternate numbering
684 * systems such as the Hebrew numbering system, or for Roman Numerals, etc.
685 * NOTE: If you use URBNF_NUMBERING_SYSTEM, you must also call setDefaultRuleSet() to
686 * specify the exact numbering system you want to use. If you want the default numbering system
687 * for the locale, call NumberFormat::createInstance() instead of creating a RuleBasedNumberFormat directly.
688 * @param locale The locale for the formatter.
689 * @param status The status indicating whether the constructor succeeded.
690 * @stable ICU 2.0
691 */
692 RuleBasedNumberFormat(URBNFRuleSetTag tag, const Locale& locale, UErrorCode& status);
693
694 //-----------------------------------------------------------------------
695 // boilerplate
696 //-----------------------------------------------------------------------
697
698 /**
699 * Copy constructor
700 * @param rhs the object to be copied from.
701 * @stable ICU 2.6
702 */
703 RuleBasedNumberFormat(const RuleBasedNumberFormat& rhs);
704
705 /**
706 * Assignment operator
707 * @param rhs the object to be copied from.
708 * @stable ICU 2.6
709 */
710 RuleBasedNumberFormat& operator=(const RuleBasedNumberFormat& rhs);
711
712 /**
713 * Release memory allocated for a RuleBasedNumberFormat when you are finished with it.
714 * @stable ICU 2.6
715 */
716 virtual ~RuleBasedNumberFormat();
717
718 /**
719 * Clone this object polymorphically. The caller is responsible
720 * for deleting the result when done.
721 * @return A copy of the object.
722 * @stable ICU 2.6
723 */
724 virtual RuleBasedNumberFormat* clone() const override;
725
726 /**
727 * Return true if the given Format objects are semantically equal.
728 * Objects of different subclasses are considered unequal.
729 * @param other the object to be compared with.
730 * @return true if the given Format objects are semantically equal.
731 * @stable ICU 2.6
732 */
733 virtual bool operator==(const Format& other) const override;
734
735 //-----------------------------------------------------------------------
736 // public API functions
737 //-----------------------------------------------------------------------
738
739 /**
740 * return the rules that were provided to the RuleBasedNumberFormat.
741 * @return the result String that was passed in
742 * @stable ICU 2.0
743 */
744 virtual UnicodeString getRules() const;
745
746 /**
747 * Return the number of public rule set names.
748 * @return the number of public rule set names.
749 * @stable ICU 2.0
750 */
751 virtual int32_t getNumberOfRuleSetNames() const;
752
753 /**
754 * Return the name of the index'th public ruleSet. If index is not valid,
755 * the function returns null.
756 * @param index the index of the ruleset
757 * @return the name of the index'th public ruleSet.
758 * @stable ICU 2.0
759 */
760 virtual UnicodeString getRuleSetName(int32_t index) const;
761
762 /**
763 * Return the number of locales for which we have localized rule set display names.
764 * @return the number of locales for which we have localized rule set display names.
765 * @stable ICU 3.2
766 */
767 virtual int32_t getNumberOfRuleSetDisplayNameLocales(void) const;
768
769 /**
770 * Return the index'th display name locale.
771 * @param index the index of the locale
772 * @param status set to a failure code when this function fails
773 * @return the locale
774 * @see #getNumberOfRuleSetDisplayNameLocales
775 * @stable ICU 3.2
776 */
777 virtual Locale getRuleSetDisplayNameLocale(int32_t index, UErrorCode& status) const;
778
779 /**
780 * Return the rule set display names for the provided locale. These are in the same order
781 * as those returned by getRuleSetName. The locale is matched against the locales for
782 * which there is display name data, using normal fallback rules. If no locale matches,
783 * the default display names are returned. (These are the internal rule set names minus
784 * the leading '%'.)
785 * @param index the index of the rule set
786 * @param locale the locale (returned by getRuleSetDisplayNameLocales) for which the localized
787 * display name is desired
788 * @return the display name for the given index, which might be bogus if there is an error
789 * @see #getRuleSetName
790 * @stable ICU 3.2
791 */
792 virtual UnicodeString getRuleSetDisplayName(int32_t index,
793 const Locale& locale = Locale::getDefault());
794
795 /**
796 * Return the rule set display name for the provided rule set and locale.
797 * The locale is matched against the locales for which there is display name data, using
798 * normal fallback rules. If no locale matches, the default display name is returned.
799 * @return the display name for the rule set
800 * @stable ICU 3.2
801 * @see #getRuleSetDisplayName
802 */
803 virtual UnicodeString getRuleSetDisplayName(const UnicodeString& ruleSetName,
804 const Locale& locale = Locale::getDefault());
805
806
807 using NumberFormat::format;
808
809 /**
810 * Formats the specified 32-bit number using the default ruleset.
811 * @param number The number to format.
812 * @param toAppendTo the string that will hold the (appended) result
813 * @param pos the fieldposition
814 * @return A textual representation of the number.
815 * @stable ICU 2.0
816 */
817 virtual UnicodeString& format(int32_t number,
818 UnicodeString& toAppendTo,
819 FieldPosition& pos) const override;
820
821 /**
822 * Formats the specified 64-bit number using the default ruleset.
823 * @param number The number to format.
824 * @param toAppendTo the string that will hold the (appended) result
825 * @param pos the fieldposition
826 * @return A textual representation of the number.
827 * @stable ICU 2.1
828 */
829 virtual UnicodeString& format(int64_t number,
830 UnicodeString& toAppendTo,
831 FieldPosition& pos) const override;
832 /**
833 * Formats the specified number using the default ruleset.
834 * @param number The number to format.
835 * @param toAppendTo the string that will hold the (appended) result
836 * @param pos the fieldposition
837 * @return A textual representation of the number.
838 * @stable ICU 2.0
839 */
840 virtual UnicodeString& format(double number,
841 UnicodeString& toAppendTo,
842 FieldPosition& pos) const override;
843
844 /**
845 * Formats the specified number using the named ruleset.
846 * @param number The number to format.
847 * @param ruleSetName The name of the rule set to format the number with.
848 * This must be the name of a valid public rule set for this formatter.
849 * @param toAppendTo the string that will hold the (appended) result
850 * @param pos the fieldposition
851 * @param status the status
852 * @return A textual representation of the number.
853 * @stable ICU 2.0
854 */
855 virtual UnicodeString& format(int32_t number,
856 const UnicodeString& ruleSetName,
857 UnicodeString& toAppendTo,
858 FieldPosition& pos,
859 UErrorCode& status) const;
860 /**
861 * Formats the specified 64-bit number using the named ruleset.
862 * @param number The number to format.
863 * @param ruleSetName The name of the rule set to format the number with.
864 * This must be the name of a valid public rule set for this formatter.
865 * @param toAppendTo the string that will hold the (appended) result
866 * @param pos the fieldposition
867 * @param status the status
868 * @return A textual representation of the number.
869 * @stable ICU 2.1
870 */
871 virtual UnicodeString& format(int64_t number,
872 const UnicodeString& ruleSetName,
873 UnicodeString& toAppendTo,
874 FieldPosition& pos,
875 UErrorCode& status) const;
876 /**
877 * Formats the specified number using the named ruleset.
878 * @param number The number to format.
879 * @param ruleSetName The name of the rule set to format the number with.
880 * This must be the name of a valid public rule set for this formatter.
881 * @param toAppendTo the string that will hold the (appended) result
882 * @param pos the fieldposition
883 * @param status the status
884 * @return A textual representation of the number.
885 * @stable ICU 2.0
886 */
887 virtual UnicodeString& format(double number,
888 const UnicodeString& ruleSetName,
889 UnicodeString& toAppendTo,
890 FieldPosition& pos,
891 UErrorCode& status) const;
892
893 protected:
894 /**
895 * Format a decimal number.
896 * The number is a DigitList wrapper onto a floating point decimal number.
897 * The default implementation in NumberFormat converts the decimal number
898 * to a double and formats that. Subclasses of NumberFormat that want
899 * to specifically handle big decimal numbers must override this method.
900 * class DecimalFormat does so.
901 *
902 * @param number The number, a DigitList format Decimal Floating Point.
903 * @param appendTo Output parameter to receive result.
904 * Result is appended to existing contents.
905 * @param pos On input: an alignment field, if desired.
906 * On output: the offsets of the alignment field.
907 * @param status Output param filled with success/failure status.
908 * @return Reference to 'appendTo' parameter.
909 * @internal
910 */
911 virtual UnicodeString& format(const number::impl::DecimalQuantity &number,
912 UnicodeString& appendTo,
913 FieldPosition& pos,
914 UErrorCode& status) const override;
915 public:
916
917 using NumberFormat::parse;
918
919 /**
920 * Parses the specified string, beginning at the specified position, according
921 * to this formatter's rules. This will match the string against all of the
922 * formatter's public rule sets and return the value corresponding to the longest
923 * parseable substring. This function's behavior is affected by the lenient
924 * parse mode.
925 * @param text The string to parse
926 * @param result the result of the parse, either a double or a long.
927 * @param parsePosition On entry, contains the position of the first character
928 * in "text" to examine. On exit, has been updated to contain the position
929 * of the first character in "text" that wasn't consumed by the parse.
930 * @see #setLenient
931 * @stable ICU 2.0
932 */
933 virtual void parse(const UnicodeString& text,
934 Formattable& result,
935 ParsePosition& parsePosition) const override;
936
937 #if !UCONFIG_NO_COLLATION
938
939 /**
940 * Turns lenient parse mode on and off.
941 *
942 * When in lenient parse mode, the formatter uses a Collator for parsing the text.
943 * Only primary differences are treated as significant. This means that case
944 * differences, accent differences, alternate spellings of the same letter
945 * (e.g., ae and a-umlaut in German), ignorable characters, etc. are ignored in
946 * matching the text. In many cases, numerals will be accepted in place of words
947 * or phrases as well.
948 *
949 * For example, all of the following will correctly parse as 255 in English in
950 * lenient-parse mode:
951 * <br>"two hundred fifty-five"
952 * <br>"two hundred fifty five"
953 * <br>"TWO HUNDRED FIFTY-FIVE"
954 * <br>"twohundredfiftyfive"
955 * <br>"2 hundred fifty-5"
956 *
957 * The Collator used is determined by the locale that was
958 * passed to this object on construction. The description passed to this object
959 * on construction may supply additional collation rules that are appended to the
960 * end of the default collator for the locale, enabling additional equivalences
961 * (such as adding more ignorable characters or permitting spelled-out version of
962 * symbols; see the demo program for examples).
963 *
964 * It's important to emphasize that even strict parsing is relatively lenient: it
965 * will accept some text that it won't produce as output. In English, for example,
966 * it will correctly parse "two hundred zero" and "fifteen hundred".
967 *
968 * @param enabled If true, turns lenient-parse mode on; if false, turns it off.
969 * @see RuleBasedCollator
970 * @stable ICU 2.0
971 */
972 virtual void setLenient(UBool enabled) override;
973
974 /**
975 * Returns true if lenient-parse mode is turned on. Lenient parsing is off
976 * by default.
977 * @return true if lenient-parse mode is turned on.
978 * @see #setLenient
979 * @stable ICU 2.0
980 */
981 virtual inline UBool isLenient(void) const override;
982
983 #endif
984
985 /**
986 * Override the default rule set to use. If ruleSetName is null, reset
987 * to the initial default rule set. If the rule set is not a public rule set name,
988 * U_ILLEGAL_ARGUMENT_ERROR is returned in status.
989 * @param ruleSetName the name of the rule set, or null to reset the initial default.
990 * @param status set to failure code when a problem occurs.
991 * @stable ICU 2.6
992 */
993 virtual void setDefaultRuleSet(const UnicodeString& ruleSetName, UErrorCode& status);
994
995 /**
996 * Return the name of the current default rule set. If the current rule set is
997 * not public, returns a bogus (and empty) UnicodeString.
998 * @return the name of the current default rule set
999 * @stable ICU 3.0
1000 */
1001 virtual UnicodeString getDefaultRuleSetName() const;
1002
1003 /**
1004 * Set a particular UDisplayContext value in the formatter, such as
1005 * UDISPCTX_CAPITALIZATION_FOR_STANDALONE. Note: For getContext, see
1006 * NumberFormat.
1007 * @param value The UDisplayContext value to set.
1008 * @param status Input/output status. If at entry this indicates a failure
1009 * status, the function will do nothing; otherwise this will be
1010 * updated with any new status from the function.
1011 * @stable ICU 53
1012 */
1013 virtual void setContext(UDisplayContext value, UErrorCode& status) override;
1014
1015 /**
1016 * Get the rounding mode.
1017 * @return A rounding mode
1018 * @stable ICU 60
1019 */
1020 virtual ERoundingMode getRoundingMode(void) const override;
1021
1022 /**
1023 * Set the rounding mode.
1024 * @param roundingMode A rounding mode
1025 * @stable ICU 60
1026 */
1027 virtual void setRoundingMode(ERoundingMode roundingMode) override;
1028
1029 public:
1030 /**
1031 * ICU "poor man's RTTI", returns a UClassID for this class.
1032 *
1033 * @stable ICU 2.8
1034 */
1035 static UClassID U_EXPORT2 getStaticClassID(void);
1036
1037 /**
1038 * ICU "poor man's RTTI", returns a UClassID for the actual class.
1039 *
1040 * @stable ICU 2.8
1041 */
1042 virtual UClassID getDynamicClassID(void) const override;
1043
1044 /**
1045 * Sets the decimal format symbols, which is generally not changed
1046 * by the programmer or user. The formatter takes ownership of
1047 * symbolsToAdopt; the client must not delete it.
1048 *
1049 * @param symbolsToAdopt DecimalFormatSymbols to be adopted.
1050 * @stable ICU 49
1051 */
1052 virtual void adoptDecimalFormatSymbols(DecimalFormatSymbols* symbolsToAdopt);
1053
1054 /**
1055 * Sets the decimal format symbols, which is generally not changed
1056 * by the programmer or user. A clone of the symbols is created and
1057 * the symbols is _not_ adopted; the client is still responsible for
1058 * deleting it.
1059 *
1060 * @param symbols DecimalFormatSymbols.
1061 * @stable ICU 49
1062 */
1063 virtual void setDecimalFormatSymbols(const DecimalFormatSymbols& symbols);
1064
1065 private:
1066 RuleBasedNumberFormat() = delete; // default constructor not implemented
1067
1068 // this will ref the localizations if they are not nullptr
1069 // caller must deref to get adoption
1070 RuleBasedNumberFormat(const UnicodeString& description, LocalizationInfo* localizations,
1071 const Locale& locale, UParseError& perror, UErrorCode& status);
1072
1073 void init(const UnicodeString& rules, LocalizationInfo* localizations, UParseError& perror, UErrorCode& status);
1074 void initCapitalizationContextInfo(const Locale& thelocale);
1075 void dispose();
1076 void stripWhitespace(UnicodeString& src);
1077 void initDefaultRuleSet();
1078 NFRuleSet* findRuleSet(const UnicodeString& name, UErrorCode& status) const;
1079
1080 /* friend access */
1081 friend class NFSubstitution;
1082 friend class NFRule;
1083 friend class NFRuleSet;
1084 friend class FractionalPartSubstitution;
1085
1086 inline NFRuleSet * getDefaultRuleSet() const;
1087 const RuleBasedCollator * getCollator() const;
1088 DecimalFormatSymbols * initializeDecimalFormatSymbols(UErrorCode &status);
1089 const DecimalFormatSymbols * getDecimalFormatSymbols() const;
1090 NFRule * initializeDefaultInfinityRule(UErrorCode &status);
1091 const NFRule * getDefaultInfinityRule() const;
1092 NFRule * initializeDefaultNaNRule(UErrorCode &status);
1093 const NFRule * getDefaultNaNRule() const;
1094 PluralFormat *createPluralFormat(UPluralType pluralType, const UnicodeString &pattern, UErrorCode& status) const;
1095 UnicodeString& adjustForCapitalizationContext(int32_t startPos, UnicodeString& currentResult, UErrorCode& status) const;
1096 UnicodeString& format(int64_t number, NFRuleSet *ruleSet, UnicodeString& toAppendTo, UErrorCode& status) const;
1097 void format(double number, NFRuleSet& rs, UnicodeString& toAppendTo, UErrorCode& status) const;
1098
1099 private:
1100 NFRuleSet **fRuleSets;
1101 UnicodeString* ruleSetDescriptions;
1102 int32_t numRuleSets;
1103 NFRuleSet *defaultRuleSet;
1104 Locale locale;
1105 RuleBasedCollator* collator;
1106 DecimalFormatSymbols* decimalFormatSymbols;
1107 NFRule *defaultInfinityRule;
1108 NFRule *defaultNaNRule;
1109 ERoundingMode fRoundingMode;
1110 UBool lenient;
1111 UnicodeString* lenientParseRules;
1112 LocalizationInfo* localizations;
1113 UnicodeString originalDescription;
1114 UBool capitalizationInfoSet;
1115 UBool capitalizationForUIListMenu;
1116 UBool capitalizationForStandAlone;
1117 BreakIterator* capitalizationBrkIter;
1118 };
1119
1120 // ---------------
1121
1122 #if !UCONFIG_NO_COLLATION
1123
1124 inline UBool
isLenient(void)1125 RuleBasedNumberFormat::isLenient(void) const {
1126 return lenient;
1127 }
1128
1129 #endif
1130
1131 inline NFRuleSet*
getDefaultRuleSet()1132 RuleBasedNumberFormat::getDefaultRuleSet() const {
1133 return defaultRuleSet;
1134 }
1135
1136 U_NAMESPACE_END
1137
1138 /* U_HAVE_RBNF */
1139 #endif
1140
1141 #endif /* U_SHOW_CPLUSPLUS_API */
1142
1143 /* RBNF_H */
1144 #endif
1145