icuSources/i18n/number_skeletons.h

   1 // © 2018 and later: Unicode, Inc. and others.
   2 // License & terms of use: http://www.unicode.org/copyright.html
   3
   4 #include "unicode/utypes.h"
   5
   6 #if !UCONFIG_NO_FORMATTING
   7 #ifndef __SOURCE_NUMBER_SKELETONS_H__
   8 #define __SOURCE_NUMBER_SKELETONS_H__
   9
  10 #include "number_types.h"
  11 #include "numparse_types.h"
  12 #include "unicode/ucharstrie.h"
  13
  14 using icu::numparse::impl::StringSegment;
  15
  16 U_NAMESPACE_BEGIN namespace number {
  17 namespace impl {
  18
  19 // Forward-declaration
  20 struct SeenMacroProps;
  21
  22 // namespace for enums and entrypoint functions
  23 namespace skeleton {
  24
  25 ///////////////////////////////////////////////////////////////////////////////////////
  26 // NOTE: For an example of how to add a new stem to the number skeleton parser, see: //
  27 // http://bugs.icu-project.org/trac/changeset/41193                                  //
  28 ///////////////////////////////////////////////////////////////////////////////////////
  29
  30 /**
  31  * While parsing a skeleton, this enum records what type of option we expect to find next.
  32  */
  33 enum ParseState {
  34
  35     // Section 0: We expect whitespace or a stem, but not an option:
  36
  37     STATE_NULL,
  38
  39     // Section 1: We might accept an option, but it is not required:
  40
  41     STATE_SCIENTIFIC,
  42     STATE_FRACTION_PRECISION,
  43
  44     // Section 2: An option is required:
  45
  46     STATE_INCREMENT_PRECISION,
  47     STATE_MEASURE_UNIT,
  48     STATE_PER_MEASURE_UNIT,
  49     STATE_CURRENCY_UNIT,
  50     STATE_INTEGER_WIDTH,
  51     STATE_NUMBERING_SYSTEM,
  52     STATE_SCALE,
  53 };
  54
  55 /**
  56  * All possible stem literals have an entry in the StemEnum. The enum name is the kebab case stem
  57  * string literal written in upper snake case.
  58  *
  59  * @see StemToObject
  60  * @see #SERIALIZED_STEM_TRIE
  61  */
  62 enum StemEnum {
  63
  64     // Section 1: Stems that do not require an option:
  65
  66     STEM_COMPACT_SHORT,
  67     STEM_COMPACT_LONG,
  68     STEM_SCIENTIFIC,
  69     STEM_ENGINEERING,
  70     STEM_NOTATION_SIMPLE,
  71     STEM_BASE_UNIT,
  72     STEM_PERCENT,
  73     STEM_PERMILLE,
  74     STEM_PRECISION_INTEGER,
  75     STEM_PRECISION_UNLIMITED,
  76     STEM_PRECISION_CURRENCY_STANDARD,
  77     STEM_PRECISION_CURRENCY_CASH,
  78     STEM_ROUNDING_MODE_CEILING,
  79     STEM_ROUNDING_MODE_FLOOR,
  80     STEM_ROUNDING_MODE_DOWN,
  81     STEM_ROUNDING_MODE_UP,
  82     STEM_ROUNDING_MODE_HALF_EVEN,
  83     STEM_ROUNDING_MODE_HALF_DOWN,
  84     STEM_ROUNDING_MODE_HALF_UP,
  85     STEM_ROUNDING_MODE_UNNECESSARY,
  86     STEM_GROUP_OFF,
  87     STEM_GROUP_MIN2,
  88     STEM_GROUP_AUTO,
  89     STEM_GROUP_ON_ALIGNED,
  90     STEM_GROUP_THOUSANDS,
  91     STEM_LATIN,
  92     STEM_UNIT_WIDTH_NARROW,
  93     STEM_UNIT_WIDTH_SHORT,
  94     STEM_UNIT_WIDTH_FULL_NAME,
  95     STEM_UNIT_WIDTH_ISO_CODE,
  96     STEM_UNIT_WIDTH_HIDDEN,
  97     STEM_SIGN_AUTO,
  98     STEM_SIGN_ALWAYS,
  99     STEM_SIGN_NEVER,
 100     STEM_SIGN_ACCOUNTING,
 101     STEM_SIGN_ACCOUNTING_ALWAYS,
 102     STEM_SIGN_EXCEPT_ZERO,
 103     STEM_SIGN_ACCOUNTING_EXCEPT_ZERO,
 104     STEM_DECIMAL_AUTO,
 105     STEM_DECIMAL_ALWAYS,
 106
 107     // Section 2: Stems that DO require an option:
 108
 109     STEM_PRECISION_INCREMENT,
 110     STEM_MEASURE_UNIT,
 111     STEM_PER_MEASURE_UNIT,
 112     STEM_CURRENCY,
 113     STEM_INTEGER_WIDTH,
 114     STEM_NUMBERING_SYSTEM,
 115     STEM_SCALE,
 116 };
 117
 118 /**
 119  * Creates a NumberFormatter corresponding to the given skeleton string.
 120  *
 121  * @param skeletonString
 122  *            A number skeleton string, possibly not in its shortest form.
 123  * @return An UnlocalizedNumberFormatter with behavior defined by the given skeleton string.
 124  */
 125 UnlocalizedNumberFormatter create(const UnicodeString& skeletonString, UErrorCode& status);
 126
 127 /**
 128  * Create a skeleton string corresponding to the given NumberFormatter.
 129  *
 130  * @param macros
 131  *            The NumberFormatter options object.
 132  * @return A skeleton string in normalized form.
 133  */
 134 UnicodeString generate(const MacroProps& macros, UErrorCode& status);
 135
 136 /**
 137  * Converts from a skeleton string to a MacroProps. This method contains the primary parse loop.
 138  *
 139  * Internal: use the create() endpoint instead of this function.
 140  */
 141 MacroProps parseSkeleton(const UnicodeString& skeletonString, UErrorCode& status);
 142
 143 /**
 144  * Given that the current segment represents a stem, parse it and save the result.
 145  *
 146  * @return The next state after parsing this stem, corresponding to what subset of options to expect.
 147  */
 148 ParseState parseStem(const StringSegment& segment, const UCharsTrie& stemTrie, SeenMacroProps& seen,
 149                      MacroProps& macros, UErrorCode& status);
 150
 151 /**
 152  * Given that the current segment represents an option, parse it and save the result.
 153  *
 154  * @return The next state after parsing this option, corresponding to what subset of options to
 155  *         expect next.
 156  */
 157 ParseState
 158 parseOption(ParseState stem, const StringSegment& segment, MacroProps& macros, UErrorCode& status);
 159
 160 } // namespace skeleton
 161
 162
 163 /**
 164  * Namespace for utility methods that convert from StemEnum to corresponding objects or enums. This
 165  * applies to only the "Section 1" stems, those that are well-defined without an option.
 166  */
 167 namespace stem_to_object {
 168
 169 Notation notation(skeleton::StemEnum stem);
 170
 171 MeasureUnit unit(skeleton::StemEnum stem);
 172
 173 Precision precision(skeleton::StemEnum stem);
 174
 175 UNumberFormatRoundingMode roundingMode(skeleton::StemEnum stem);
 176
 177 UGroupingStrategy groupingStrategy(skeleton::StemEnum stem);
 178
 179 UNumberUnitWidth unitWidth(skeleton::StemEnum stem);
 180
 181 UNumberSignDisplay signDisplay(skeleton::StemEnum stem);
 182
 183 UNumberDecimalSeparatorDisplay decimalSeparatorDisplay(skeleton::StemEnum stem);
 184
 185 } // namespace stem_to_object
 186
 187 /**
 188  * Namespace for utility methods that convert from enums to stem strings. More complex object conversions
 189  * take place in the object_to_stem_string namespace.
 190  */
 191 namespace enum_to_stem_string {
 192
 193 void roundingMode(UNumberFormatRoundingMode value, UnicodeString& sb);
 194
 195 void groupingStrategy(UGroupingStrategy value, UnicodeString& sb);
 196
 197 void unitWidth(UNumberUnitWidth value, UnicodeString& sb);
 198
 199 void signDisplay(UNumberSignDisplay value, UnicodeString& sb);
 200
 201 void decimalSeparatorDisplay(UNumberDecimalSeparatorDisplay value, UnicodeString& sb);
 202
 203 } // namespace enum_to_stem_string
 204
 205 /**
 206  * Namespace for utility methods for processing stems and options that cannot be interpreted literally.
 207  */
 208 namespace blueprint_helpers {
 209
 210 /** @return Whether we successfully found and parsed an exponent width option. */
 211 bool parseExponentWidthOption(const StringSegment& segment, MacroProps& macros, UErrorCode& status);
 212
 213 void generateExponentWidthOption(int32_t minExponentDigits, UnicodeString& sb, UErrorCode& status);
 214
 215 /** @return Whether we successfully found and parsed an exponent sign option. */
 216 bool parseExponentSignOption(const StringSegment& segment, MacroProps& macros, UErrorCode& status);
 217
 218 void parseCurrencyOption(const StringSegment& segment, MacroProps& macros, UErrorCode& status);
 219
 220 void generateCurrencyOption(const CurrencyUnit& currency, UnicodeString& sb, UErrorCode& status);
 221
 222 void parseMeasureUnitOption(const StringSegment& segment, MacroProps& macros, UErrorCode& status);
 223
 224 void generateMeasureUnitOption(const MeasureUnit& measureUnit, UnicodeString& sb, UErrorCode& status);
 225
 226 void parseMeasurePerUnitOption(const StringSegment& segment, MacroProps& macros, UErrorCode& status);
 227
 228 void parseFractionStem(const StringSegment& segment, MacroProps& macros, UErrorCode& status);
 229
 230 void generateFractionStem(int32_t minFrac, int32_t maxFrac, UnicodeString& sb, UErrorCode& status);
 231
 232 void parseDigitsStem(const StringSegment& segment, MacroProps& macros, UErrorCode& status);
 233
 234 void generateDigitsStem(int32_t minSig, int32_t maxSig, UnicodeString& sb, UErrorCode& status);
 235
 236 /** @return Whether we successfully found and parsed a frac-sig option. */
 237 bool parseFracSigOption(const StringSegment& segment, MacroProps& macros, UErrorCode& status);
 238
 239 void parseIncrementOption(const StringSegment& segment, MacroProps& macros, UErrorCode& status);
 240
 241 void
 242 generateIncrementOption(double increment, int32_t trailingZeros, UnicodeString& sb, UErrorCode& status);
 243
 244 void parseIntegerWidthOption(const StringSegment& segment, MacroProps& macros, UErrorCode& status);
 245
 246 void generateIntegerWidthOption(int32_t minInt, int32_t maxInt, UnicodeString& sb, UErrorCode& status);
 247
 248 void parseNumberingSystemOption(const StringSegment& segment, MacroProps& macros, UErrorCode& status);
 249
 250 void generateNumberingSystemOption(const NumberingSystem& ns, UnicodeString& sb, UErrorCode& status);
 251
 252 void parseScaleOption(const StringSegment& segment, MacroProps& macros, UErrorCode& status);
 253
 254 void generateScaleOption(int32_t magnitude, const DecNum* arbitrary, UnicodeString& sb,
 255                               UErrorCode& status);
 256
 257 } // namespace blueprint_helpers
 258
 259 /**
 260  * Class for utility methods for generating a token corresponding to each macro-prop. Each method
 261  * returns whether or not a token was written to the string builder.
 262  *
 263  * This needs to be a class, not a namespace, so it can be friended.
 264  */
 265 class GeneratorHelpers {
 266   public:
 267     /**
 268      * Main skeleton generator function. Appends the normalized skeleton for the MacroProps to the given
 269      * StringBuilder.
 270      *
 271      * Internal: use the create() endpoint instead of this function.
 272      */
 273     static void generateSkeleton(const MacroProps& macros, UnicodeString& sb, UErrorCode& status);
 274
 275   private:
 276     static bool notation(const MacroProps& macros, UnicodeString& sb, UErrorCode& status);
 277
 278     static bool unit(const MacroProps& macros, UnicodeString& sb, UErrorCode& status);
 279
 280     static bool perUnit(const MacroProps& macros, UnicodeString& sb, UErrorCode& status);
 281
 282     static bool precision(const MacroProps& macros, UnicodeString& sb, UErrorCode& status);
 283
 284     static bool roundingMode(const MacroProps& macros, UnicodeString& sb, UErrorCode& status);
 285
 286     static bool grouping(const MacroProps& macros, UnicodeString& sb, UErrorCode& status);
 287
 288     static bool integerWidth(const MacroProps& macros, UnicodeString& sb, UErrorCode& status);
 289
 290     static bool symbols(const MacroProps& macros, UnicodeString& sb, UErrorCode& status);
 291
 292     static bool unitWidth(const MacroProps& macros, UnicodeString& sb, UErrorCode& status);
 293
 294     static bool sign(const MacroProps& macros, UnicodeString& sb, UErrorCode& status);
 295
 296     static bool decimal(const MacroProps& macros, UnicodeString& sb, UErrorCode& status);
 297
 298     static bool scale(const MacroProps& macros, UnicodeString& sb, UErrorCode& status);
 299
 300 };
 301
 302 /**
 303  * Struct for null-checking.
 304  * In Java, we can just check the object reference. In C++, we need a different method.
 305  */
 306 struct SeenMacroProps {
 307     bool notation = false;
 308     bool unit = false;
 309     bool perUnit = false;
 310     bool precision = false;
 311     bool roundingMode = false;
 312     bool grouper = false;
 313     bool padder = false;
 314     bool integerWidth = false;
 315     bool symbols = false;
 316     bool unitWidth = false;
 317     bool sign = false;
 318     bool decimal = false;
 319     bool scale = false;
 320 };
 321
 322 } // namespace impl
 323 } // namespace number
 324 U_NAMESPACE_END
 325
 326 #endif //__SOURCE_NUMBER_SKELETONS_H__
 327 #endif /* #if !UCONFIG_NO_FORMATTING */