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