#include "unicode/utypes.h"
#if !UCONFIG_NO_BREAK_ITERATION
#include "brkeng.h"
#include "dictbe.h"
#include "unicode/uniset.h"
#include "unicode/chariter.h"
#include "unicode/ubrk.h"
#include "uvector.h"
#include "triedict.h"
U_NAMESPACE_BEGIN
DictionaryBreakEngine::DictionaryBreakEngine(uint32_t breakTypes) {
fTypes = breakTypes;
}
DictionaryBreakEngine::~DictionaryBreakEngine() {
}
UBool
DictionaryBreakEngine::handles(UChar32 c, int32_t breakType) const {
return (breakType >= 0 && breakType < 32 && (((uint32_t)1 << breakType) & fTypes)
&& fSet.contains(c));
}
int32_t
DictionaryBreakEngine::findBreaks( UText *text,
int32_t startPos,
int32_t endPos,
UBool reverse,
int32_t breakType,
UStack &foundBreaks ) const {
int32_t result = 0;
int32_t start = (int32_t)utext_getNativeIndex(text);
int32_t current;
int32_t rangeStart;
int32_t rangeEnd;
UChar32 c = utext_current32(text);
if (reverse) {
UBool isDict = fSet.contains(c);
while((current = (int32_t)utext_getNativeIndex(text)) > startPos && isDict) {
c = utext_previous32(text);
isDict = fSet.contains(c);
}
rangeStart = (current < startPos) ? startPos : current+(isDict ? 0 : 1);
rangeEnd = start + 1;
}
else {
while((current = (int32_t)utext_getNativeIndex(text)) < endPos && fSet.contains(c)) {
utext_next32(text); c = utext_current32(text);
}
rangeStart = start;
rangeEnd = current;
}
if (breakType >= 0 && breakType < 32 && (((uint32_t)1 << breakType) & fTypes)) {
result = divideUpDictionaryRange(text, rangeStart, rangeEnd, foundBreaks);
utext_setNativeIndex(text, current);
}
return result;
}
void
DictionaryBreakEngine::setCharacters( const UnicodeSet &set ) {
fSet = set;
fSet.compact();
}
#define POSSIBLE_WORD_LIST_MAX 20
class PossibleWord {
private:
int32_t lengths[POSSIBLE_WORD_LIST_MAX];
int count; int32_t prefix; int32_t offset; int mark; int current;
public:
PossibleWord();
~PossibleWord();
int candidates( UText *text, const TrieWordDictionary *dict, int32_t rangeEnd );
int32_t acceptMarked( UText *text );
UBool backUp( UText *text );
int32_t longestPrefix();
void markCurrent();
};
inline
PossibleWord::PossibleWord() {
offset = -1;
}
inline
PossibleWord::~PossibleWord() {
}
inline int
PossibleWord::candidates( UText *text, const TrieWordDictionary *dict, int32_t rangeEnd ) {
int32_t start = (int32_t)utext_getNativeIndex(text);
if (start != offset) {
offset = start;
prefix = dict->matches(text, rangeEnd-start, lengths, count, sizeof(lengths)/sizeof(lengths[0]));
if (count <= 0) {
utext_setNativeIndex(text, start);
}
}
if (count > 0) {
utext_setNativeIndex(text, start+lengths[count-1]);
}
current = count-1;
mark = current;
return count;
}
inline int32_t
PossibleWord::acceptMarked( UText *text ) {
utext_setNativeIndex(text, offset + lengths[mark]);
return lengths[mark];
}
inline UBool
PossibleWord::backUp( UText *text ) {
if (current > 0) {
utext_setNativeIndex(text, offset + lengths[--current]);
return TRUE;
}
return FALSE;
}
inline int32_t
PossibleWord::longestPrefix() {
return prefix;
}
inline void
PossibleWord::markCurrent() {
mark = current;
}
#define THAI_LOOKAHEAD 3
#define THAI_ROOT_COMBINE_THRESHOLD 3
#define THAI_PREFIX_COMBINE_THRESHOLD 3
#define THAI_PAIYANNOI 0x0E2F
#define THAI_MAIYAMOK 0x0E46
#define THAI_MIN_WORD 2
#define THAI_MIN_WORD_SPAN (THAI_MIN_WORD * 2)
ThaiBreakEngine::ThaiBreakEngine(const TrieWordDictionary *adoptDictionary, UErrorCode &status)
: DictionaryBreakEngine((1<<UBRK_WORD) | (1<<UBRK_LINE)),
fDictionary(adoptDictionary)
{
fThaiWordSet.applyPattern(UNICODE_STRING_SIMPLE("[[:Thai:]&[:LineBreak=SA:]]"), status);
if (U_SUCCESS(status)) {
setCharacters(fThaiWordSet);
}
fMarkSet.applyPattern(UNICODE_STRING_SIMPLE("[[:Thai:]&[:LineBreak=SA:]&[:M:]]"), status);
fMarkSet.add(0x0020);
fEndWordSet = fThaiWordSet;
fEndWordSet.remove(0x0E31); fEndWordSet.remove(0x0E40, 0x0E44); fBeginWordSet.add(0x0E01, 0x0E2E); fBeginWordSet.add(0x0E40, 0x0E44); fSuffixSet.add(THAI_PAIYANNOI);
fSuffixSet.add(THAI_MAIYAMOK);
fMarkSet.compact();
fEndWordSet.compact();
fBeginWordSet.compact();
fSuffixSet.compact();
}
ThaiBreakEngine::~ThaiBreakEngine() {
delete fDictionary;
}
int32_t
ThaiBreakEngine::divideUpDictionaryRange( UText *text,
int32_t rangeStart,
int32_t rangeEnd,
UStack &foundBreaks ) const {
if ((rangeEnd - rangeStart) < THAI_MIN_WORD_SPAN) {
return 0; }
uint32_t wordsFound = 0;
int32_t wordLength;
int32_t current;
UErrorCode status = U_ZERO_ERROR;
PossibleWord words[THAI_LOOKAHEAD];
UChar32 uc;
utext_setNativeIndex(text, rangeStart);
while (U_SUCCESS(status) && (current = (int32_t)utext_getNativeIndex(text)) < rangeEnd) {
wordLength = 0;
int candidates = words[wordsFound%THAI_LOOKAHEAD].candidates(text, fDictionary, rangeEnd);
if (candidates == 1) {
wordLength = words[wordsFound%THAI_LOOKAHEAD].acceptMarked(text);
wordsFound += 1;
}
else if (candidates > 1) {
if ((int32_t)utext_getNativeIndex(text) >= rangeEnd) {
goto foundBest;
}
do {
int wordsMatched = 1;
if (words[(wordsFound+1)%THAI_LOOKAHEAD].candidates(text, fDictionary, rangeEnd) > 0) {
if (wordsMatched < 2) {
words[wordsFound%THAI_LOOKAHEAD].markCurrent();
wordsMatched = 2;
}
if ((int32_t)utext_getNativeIndex(text) >= rangeEnd) {
goto foundBest;
}
do {
if (words[(wordsFound+2)%THAI_LOOKAHEAD].candidates(text, fDictionary, rangeEnd)) {
words[wordsFound%THAI_LOOKAHEAD].markCurrent();
goto foundBest;
}
}
while (words[(wordsFound+1)%THAI_LOOKAHEAD].backUp(text));
}
}
while (words[wordsFound%THAI_LOOKAHEAD].backUp(text));
foundBest:
wordLength = words[wordsFound%THAI_LOOKAHEAD].acceptMarked(text);
wordsFound += 1;
}
if ((int32_t)utext_getNativeIndex(text) < rangeEnd && wordLength < THAI_ROOT_COMBINE_THRESHOLD) {
if (words[wordsFound%THAI_LOOKAHEAD].candidates(text, fDictionary, rangeEnd) <= 0
&& (wordLength == 0
|| words[wordsFound%THAI_LOOKAHEAD].longestPrefix() < THAI_PREFIX_COMBINE_THRESHOLD)) {
int32_t remaining = rangeEnd - (current+wordLength);
UChar32 pc = utext_current32(text);
int32_t chars = 0;
for (;;) {
utext_next32(text);
uc = utext_current32(text);
chars += 1;
if (--remaining <= 0) {
break;
}
if (fEndWordSet.contains(pc) && fBeginWordSet.contains(uc)) {
int candidates = words[(wordsFound+1)%THAI_LOOKAHEAD].candidates(text, fDictionary, rangeEnd);
utext_setNativeIndex(text, current+wordLength+chars);
if (candidates > 0) {
break;
}
}
pc = uc;
}
if (wordLength <= 0) {
wordsFound += 1;
}
wordLength += chars;
}
else {
utext_setNativeIndex(text, current+wordLength);
}
}
int32_t currPos;
while ((currPos = (int32_t)utext_getNativeIndex(text)) < rangeEnd && fMarkSet.contains(utext_current32(text))) {
utext_next32(text);
wordLength += (int32_t)utext_getNativeIndex(text) - currPos;
}
if ((int32_t)utext_getNativeIndex(text) < rangeEnd && wordLength > 0) {
if (words[wordsFound%THAI_LOOKAHEAD].candidates(text, fDictionary, rangeEnd) <= 0
&& fSuffixSet.contains(uc = utext_current32(text))) {
if (uc == THAI_PAIYANNOI) {
if (!fSuffixSet.contains(utext_previous32(text))) {
utext_next32(text);
utext_next32(text);
wordLength += 1; uc = utext_current32(text); }
else {
utext_next32(text);
}
}
if (uc == THAI_MAIYAMOK) {
if (utext_previous32(text) != THAI_MAIYAMOK) {
utext_next32(text);
utext_next32(text);
wordLength += 1; }
else {
utext_next32(text);
}
}
}
else {
utext_setNativeIndex(text, current+wordLength);
}
}
if (wordLength > 0) {
foundBreaks.push((current+wordLength), status);
}
}
if (foundBreaks.peeki() >= rangeEnd) {
(void) foundBreaks.popi();
wordsFound -= 1;
}
return wordsFound;
}
U_NAMESPACE_END
#endif