AuroraMiddleware/scuffed-code
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

ICU-20418 Implementing concise number skeletons in ICU4J.

Browse Source
This commit is contained in:
Shane Carr 2019-12-13 17:30:31 -08:00 committed by Shane F. Carr
parent 0ad2f9590b
commit c98af924a1
5 changed files with 449 additions and 6 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
icu4j/main/classes/core/src/com/ibm/icu/impl/StringSegment.java
View File

@ -221,8 +221,14 @@ public class StringSegment implements CharSequence {
return Utility.charSequenceHashCode(this);
}
/** Returns a string representation useful for debugging. */
@Override
public String toString() {
return str.substring(0, start) + "[" + str.substring(start, end) + "]" + str.substring(end);
}
/** Returns a String that is equivalent to the CharSequence representation. */
public String asString() {
return str.substring(start, end);
}
}

126
icu4j/main/classes/core/src/com/ibm/icu/number/NumberSkeletonImpl.java
View File

@ -53,6 +53,7 @@ class NumberSkeletonImpl {
STATE_INCREMENT_PRECISION,
STATE_MEASURE_UNIT,
STATE_PER_MEASURE_UNIT,
STATE_IDENTIFIER_UNIT,
STATE_CURRENCY_UNIT,
STATE_INTEGER_WIDTH,
STATE_NUMBERING_SYSTEM,
@ -76,6 +77,7 @@ class NumberSkeletonImpl {
STEM_BASE_UNIT,
STEM_PERCENT,
STEM_PERMILLE,
STEM_PERCENT_100, // concise-only
STEM_PRECISION_INTEGER,
STEM_PRECISION_UNLIMITED,
STEM_PRECISION_CURRENCY_STANDARD,
@ -113,6 +115,7 @@ class NumberSkeletonImpl {
STEM_PRECISION_INCREMENT,
STEM_MEASURE_UNIT,
STEM_PER_MEASURE_UNIT,
STEM_UNIT,
STEM_CURRENCY,
STEM_INTEGER_WIDTH,
STEM_NUMBERING_SYSTEM,
@ -174,11 +177,27 @@ class NumberSkeletonImpl {
b.add("precision-increment", StemEnum.STEM_PRECISION_INCREMENT.ordinal());
b.add("measure-unit", StemEnum.STEM_MEASURE_UNIT.ordinal());
b.add("per-measure-unit", StemEnum.STEM_PER_MEASURE_UNIT.ordinal());
b.add("unit", StemEnum.STEM_UNIT.ordinal());
b.add("currency", StemEnum.STEM_CURRENCY.ordinal());
b.add("integer-width", StemEnum.STEM_INTEGER_WIDTH.ordinal());
b.add("numbering-system", StemEnum.STEM_NUMBERING_SYSTEM.ordinal());
b.add("scale", StemEnum.STEM_SCALE.ordinal());
// Section 3 (concise tokens):
b.add("K", StemEnum.STEM_COMPACT_SHORT.ordinal());
b.add("KK", StemEnum.STEM_COMPACT_LONG.ordinal());
b.add("%", StemEnum.STEM_PERCENT.ordinal());
b.add("%x100", StemEnum.STEM_PERCENT_100.ordinal());
b.add(",_", StemEnum.STEM_GROUP_OFF.ordinal());
b.add(",?", StemEnum.STEM_GROUP_MIN2.ordinal());
b.add(",!", StemEnum.STEM_GROUP_ON_ALIGNED.ordinal());
b.add("+!", StemEnum.STEM_SIGN_ALWAYS.ordinal());
b.add("+_", StemEnum.STEM_SIGN_NEVER.ordinal());
b.add("()", StemEnum.STEM_SIGN_ACCOUNTING.ordinal());
b.add("()!", StemEnum.STEM_SIGN_ACCOUNTING_ALWAYS.ordinal());
b.add("+?", StemEnum.STEM_SIGN_EXCEPT_ZERO.ordinal());
b.add("()?", StemEnum.STEM_SIGN_ACCOUNTING_EXCEPT_ZERO.ordinal());
// Build the CharsTrie
// TODO: Use SLOW or FAST here?
return b.buildCharSequence(StringTrieBuilder.Option.FAST).toString();
@ -604,6 +623,14 @@ class NumberSkeletonImpl {
checkNull(macros.precision, segment);
BlueprintHelpers.parseDigitsStem(segment, macros);
return ParseState.STATE_NULL;
case 'E':
checkNull(macros.notation, segment);
BlueprintHelpers.parseScientificStem(segment, macros);
return ParseState.STATE_NULL;
case '0':
checkNull(macros.notation, segment);
BlueprintHelpers.parseIntegerStem(segment, macros);
return ParseState.STATE_NULL;
}
// Now look at the stemsTrie, which is already be pointing at our stem.
@ -641,6 +668,13 @@ class NumberSkeletonImpl {
macros.unit = StemToObject.unit(stem);
return ParseState.STATE_NULL;
case STEM_PERCENT_100:
checkNull(macros.scale, segment);
checkNull(macros.unit, segment);
macros.scale = Scale.powerOfTen(2);
macros.unit = NoUnit.PERCENT;
return ParseState.STATE_NULL;
case STEM_PRECISION_INTEGER:
case STEM_PRECISION_UNLIMITED:
case STEM_PRECISION_CURRENCY_STANDARD:
@ -720,6 +754,11 @@ class NumberSkeletonImpl {
checkNull(macros.perUnit, segment);
return ParseState.STATE_PER_MEASURE_UNIT;
case STEM_UNIT:
checkNull(macros.unit, segment);
checkNull(macros.perUnit, segment);
return ParseState.STATE_IDENTIFIER_UNIT;
case STEM_CURRENCY:
checkNull(macros.unit, segment);
return ParseState.STATE_CURRENCY_UNIT;
@ -761,6 +800,9 @@ class NumberSkeletonImpl {
case STATE_PER_MEASURE_UNIT:
BlueprintHelpers.parseMeasurePerUnitOption(segment, macros);
return ParseState.STATE_NULL;
case STATE_IDENTIFIER_UNIT:
BlueprintHelpers.parseIdentifierUnitOption(segment, macros);
return ParseState.STATE_NULL;
case STATE_INCREMENT_PRECISION:
BlueprintHelpers.parseIncrementOption(segment, macros);
return ParseState.STATE_NULL;
@ -970,7 +1012,7 @@ class NumberSkeletonImpl {
}
private static void parseMeasurePerUnitOption(StringSegment segment, MacroProps macros) {
// A little bit of a hack: safe the current unit (numerator), call the main measure unit
// A little bit of a hack: save the current unit (numerator), call the main measure unit
// parsing code, put back the numerator unit, and put the new unit into per-unit.
MeasureUnit numerator = macros.unit;
parseMeasureUnitOption(segment, macros);
@ -978,6 +1020,17 @@ class NumberSkeletonImpl {
macros.unit = numerator;
}
private static void parseIdentifierUnitOption(StringSegment segment, MacroProps macros) {
MeasureUnit[] units = MeasureUnit.parseCoreUnitIdentifier(segment.asString());
if (units == null) {
throw new SkeletonSyntaxException("Invalid core unit identifier", segment);
}
macros.unit = units[0];
if (units.length == 2) {
macros.perUnit = units[1];
}
}
private static void parseFractionStem(StringSegment segment, MacroProps macros) {
assert segment.charAt(0) == '.';
int offset = 1;
@ -1012,7 +1065,11 @@ class NumberSkeletonImpl {
}
// Use the public APIs to enforce bounds checking
if (maxFrac == -1) {
macros.precision = Precision.minFraction(minFrac);
if (minFrac == 0) {
macros.precision = Precision.unlimited();
} else {
macros.precision = Precision.minFraction(minFrac);
}
} else {
macros.precision = Precision.minMaxFraction(minFrac, maxFrac);
}
@ -1081,6 +1138,71 @@ class NumberSkeletonImpl {
}
}
private static void parseScientificStem(StringSegment segment, MacroProps macros) {
assert(segment.charAt(0) == 'E');
block:
{
int offset = 1;
if (segment.length() == offset) {
break block;
}
boolean isEngineering = false;
if (segment.charAt(offset) == 'E') {
isEngineering = true;
offset++;
if (segment.length() == offset) {
break block;
}
}
SignDisplay signDisplay = SignDisplay.AUTO;
if (segment.charAt(offset) == '+') {
offset++;
if (segment.length() == offset) {
break block;
}
if (segment.charAt(offset) == '!') {
signDisplay = SignDisplay.ALWAYS;
} else if (segment.charAt(offset) == '?') {
signDisplay = SignDisplay.EXCEPT_ZERO;
} else {
break block;
}
offset++;
if (segment.length() == offset) {
break block;
}
}
int minDigits = 0;
for (; offset < segment.length(); offset++) {
if (segment.charAt(offset) != '0') {
break block;
}
minDigits++;
}
macros.notation = (isEngineering ? Notation.engineering() : Notation.scientific())
.withExponentSignDisplay(signDisplay)
.withMinExponentDigits(minDigits);
return;
}
throw new SkeletonSyntaxException("Invalid scientific stem", segment);
}
private static void parseIntegerStem(StringSegment segment, MacroProps macros) {
assert(segment.charAt(0) == '0');
int offset = 1;
for (; offset < segment.length(); offset++) {
if (segment.charAt(offset) != '0') {
offset--;
break;
}
}
if (offset < segment.length()) {
throw new SkeletonSyntaxException("Invalid integer stem", segment);
}
macros.integerWidth = IntegerWidth.zeroFillTo(offset);
return;
}
/** @return Whether we successfully found and parsed a frac-sig option. */
private static boolean parseFracSigOption(StringSegment segment, MacroProps macros) {
if (segment.charAt(0) != '@') {

41
icu4j/main/classes/core/src/com/ibm/icu/util/MeasureUnit.java
View File

@ -205,6 +205,47 @@ public class MeasureUnit implements Serializable {
return MeasureUnit.addUnit(type, subType, factory);
}
private static MeasureUnit findBySubType(String subType) {
populateCache();
for (Map<String, MeasureUnit> unitsForType : cache.values()) {
if (unitsForType.containsKey(subType)) {
return unitsForType.get(subType);
}
}
return null;
}
/**
* For ICU use only.
* @internal
* @deprecated This API is ICU internal only.
*/
@Deprecated
public static MeasureUnit[] parseCoreUnitIdentifier(String coreUnitIdentifier) {
// First search for the whole code unit identifier as a subType
MeasureUnit whole = findBySubType(coreUnitIdentifier);
if (whole != null) {
return new MeasureUnit[] { whole }; // found a numerator but not denominator
}
// If not found, try breaking apart numerator and denominator
int perIdx = coreUnitIdentifier.indexOf("-per-");
if (perIdx == -1) {
// String does not contain "-per-"
return null;
}
String numeratorStr = coreUnitIdentifier.substring(0, perIdx);
String denominatorStr = coreUnitIdentifier.substring(perIdx + 5);
MeasureUnit numerator = findBySubType(numeratorStr);
MeasureUnit denominator = findBySubType(denominatorStr);
if (numerator != null && denominator != null) {
return new MeasureUnit[] { numerator, denominator }; // found both a numerator and denominator
}
// The numerator or denominator were invalid
return null;
}
/**
* For ICU use only.
* @internal

251
icu4j/main/tests/core/src/com/ibm/icu/dev/test/number/NumberFormatterApiTest.java
View File

File diff suppressed because it is too large Load Diff

31
icu4j/main/tests/core/src/com/ibm/icu/dev/test/number/NumberSkeletonTest.java
View File

@ -57,6 +57,7 @@ public class NumberSkeletonTest {
"measure-unit/length-meter",
"measure-unit/area-square-meter",
"measure-unit/energy-joule per-measure-unit/length-meter",
"unit/square-meter-per-square-meter",
"currency/XXX",
"currency/ZZZ",
"currency/usd",
@ -92,7 +93,20 @@ public class NumberSkeletonTest {
"numbering-system/latn",
"precision-integer/@##",
"precision-integer rounding-mode-ceiling",
"precision-currency-cash rounding-mode-ceiling" };
"precision-currency-cash rounding-mode-ceiling",
"0",
"00",
"000",
"E0",
"E00",
"E000",
"EE0",
"EE00",
"EE+?0",
"EE+?00",
"EE+!0",
"EE+!00",
};
for (String cas : cases) {
try {
@ -137,7 +151,20 @@ public class NumberSkeletonTest {
"integer-width/+0#",
"integer-width/+#",
"integer-width/+#0",
"scientific/foo" };
"scientific/foo",
"E",
"E1",
"E+",
"E+?",
"E+!",
"E+0",
"EE",
"EE+",
"EEE",
"EEE0",
"001",
"00+",
};
for (String cas : cases) {
try {