diff options
Diffstat (limited to 'native/src')
| -rw-r--r-- | native/src/bigram_dictionary.cpp | 26 | ||||
| -rw-r--r-- | native/src/defines.h | 10 | ||||
| -rw-r--r-- | native/src/dictionary.cpp | 2 | ||||
| -rw-r--r-- | native/src/proximity_info.h | 6 | ||||
| -rw-r--r-- | native/src/unigram_dictionary.cpp | 207 | ||||
| -rw-r--r-- | native/src/unigram_dictionary.h | 2 |
6 files changed, 200 insertions, 53 deletions
diff --git a/native/src/bigram_dictionary.cpp b/native/src/bigram_dictionary.cpp index 5ec310f07..36761b88d 100644 --- a/native/src/bigram_dictionary.cpp +++ b/native/src/bigram_dictionary.cpp @@ -30,8 +30,10 @@ BigramDictionary::BigramDictionary(const unsigned char *dict, int maxWordLength, : DICT(dict), MAX_WORD_LENGTH(maxWordLength), MAX_ALTERNATIVES(maxAlternatives), IS_LATEST_DICT_VERSION(isLatestDictVersion), HAS_BIGRAM(hasBigram), mParentDictionary(parentDictionary) { - if (DEBUG_DICT) LOGI("BigramDictionary - constructor"); - if (DEBUG_DICT) LOGI("Has Bigram : %d", hasBigram); + if (DEBUG_DICT) { + LOGI("BigramDictionary - constructor"); + LOGI("Has Bigram : %d", hasBigram); + } } BigramDictionary::~BigramDictionary() { @@ -54,7 +56,9 @@ bool BigramDictionary::addWordBigram(unsigned short *word, int length, int frequ } insertAt++; } - if (DEBUG_DICT) LOGI("Bigram: InsertAt -> %d maxBigrams: %d", insertAt, mMaxBigrams); + if (DEBUG_DICT) { + LOGI("Bigram: InsertAt -> %d maxBigrams: %d", insertAt, mMaxBigrams); + } if (insertAt < mMaxBigrams) { memmove((char*) mBigramFreq + (insertAt + 1) * sizeof(mBigramFreq[0]), (char*) mBigramFreq + insertAt * sizeof(mBigramFreq[0]), @@ -68,7 +72,9 @@ bool BigramDictionary::addWordBigram(unsigned short *word, int length, int frequ *dest++ = *word++; } *dest = 0; // NULL terminate - if (DEBUG_DICT) LOGI("Bigram: Added word at %d", insertAt); + if (DEBUG_DICT) { + LOGI("Bigram: Added word at %d", insertAt); + } return true; } return false; @@ -107,7 +113,9 @@ int BigramDictionary::getBigrams(unsigned short *prevWord, int prevWordLength, i if (HAS_BIGRAM && IS_LATEST_DICT_VERSION) { int pos = mParentDictionary->isValidWordRec( DICTIONARY_HEADER_SIZE, prevWord, 0, prevWordLength); - if (DEBUG_DICT) LOGI("Pos -> %d", pos); + if (DEBUG_DICT) { + LOGI("Pos -> %d", pos); + } if (pos < 0) { return 0; } @@ -151,7 +159,9 @@ void BigramDictionary::searchForTerminalNode(int addressLookingFor, int frequenc } pos = followDownBranchAddress; // pos start at count int count = DICT[pos] & 0xFF; - if (DEBUG_DICT) LOGI("count - %d",count); + if (DEBUG_DICT) { + LOGI("count - %d",count); + } pos++; for (int i = 0; i < count; i++) { // pos at data @@ -225,7 +235,9 @@ void BigramDictionary::searchForTerminalNode(int addressLookingFor, int frequenc } depth++; if (followDownBranchAddress == 0) { - if (DEBUG_DICT) LOGI("ERROR!!! Cannot find bigram!!"); + if (DEBUG_DICT) { + LOGI("ERROR!!! Cannot find bigram!!"); + } break; } } diff --git a/native/src/defines.h b/native/src/defines.h index 00cbb6c9e..0a3240507 100644 --- a/native/src/defines.h +++ b/native/src/defines.h @@ -77,8 +77,8 @@ static void prof_out(void) { } #else // FLAG_DBG -#define LOGE -#define LOGI +#define LOGE(fmt, ...) +#define LOGI(fmt, ...) #define DEBUG_DICT false #define DEBUG_DICT_FULL false #define DEBUG_SHOW_FOUND_WORD false @@ -138,12 +138,14 @@ static void prof_out(void) { #define SUGGEST_WORDS_WITH_SPACE_PROXIMITY true // The following "rate"s are used as a multiplier before dividing by 100, so they are in percent. -#define WORDS_WITH_MISSING_CHARACTER_DEMOTION_RATE 70 -#define WORDS_WITH_MISSING_SPACE_CHARACTER_DEMOTION_RATE 80 +#define WORDS_WITH_MISSING_CHARACTER_DEMOTION_RATE 80 +#define WORDS_WITH_MISSING_CHARACTER_DEMOTION_START_POS_10X 12 +#define WORDS_WITH_MISSING_SPACE_CHARACTER_DEMOTION_RATE 67 #define WORDS_WITH_EXCESSIVE_CHARACTER_DEMOTION_RATE 75 #define WORDS_WITH_EXCESSIVE_CHARACTER_OUT_OF_PROXIMITY_DEMOTION_RATE 75 #define WORDS_WITH_TRANSPOSED_CHARACTERS_DEMOTION_RATE 60 #define FULL_MATCHED_WORDS_PROMOTION_RATE 120 +#define WORDS_WITH_PROXIMITY_CHARACTER_DEMOTION_RATE 90 // This should be greater than or equal to MAX_WORD_LENGTH defined in BinaryDictionary.java // This is only used for the size of array. Not to be used in c functions. diff --git a/native/src/dictionary.cpp b/native/src/dictionary.cpp index d69cb2a53..d9ef8f3b7 100644 --- a/native/src/dictionary.cpp +++ b/native/src/dictionary.cpp @@ -53,7 +53,7 @@ bool Dictionary::hasBigram() { return ((mDict[1] & 0xFF) == 1); } -// TODO: use uint16_t instead of unsigned short +// TODO: use uint32_t instead of unsigned short bool Dictionary::isValidWord(unsigned short *word, int length) { if (IS_LATEST_DICT_VERSION) { return (isValidWordRec(DICTIONARY_HEADER_SIZE, word, 0, length) != NOT_VALID_WORD); diff --git a/native/src/proximity_info.h b/native/src/proximity_info.h index 0f1201866..c2062e8c5 100644 --- a/native/src/proximity_info.h +++ b/native/src/proximity_info.h @@ -32,13 +32,13 @@ public: bool hasSpaceProximity(const int x, const int y) const; private: int getStartIndexFromCoordinates(const int x, const int y) const; - const int CELL_WIDTH; - const int CELL_HEIGHT; + const int MAX_PROXIMITY_CHARS_SIZE; const int KEYBOARD_WIDTH; const int KEYBOARD_HEIGHT; const int GRID_WIDTH; const int GRID_HEIGHT; - const int MAX_PROXIMITY_CHARS_SIZE; + const int CELL_WIDTH; + const int CELL_HEIGHT; uint32_t *mProximityCharsArray; }; }; // namespace latinime diff --git a/native/src/unigram_dictionary.cpp b/native/src/unigram_dictionary.cpp index 30fbaeae1..e49e95b81 100644 --- a/native/src/unigram_dictionary.cpp +++ b/native/src/unigram_dictionary.cpp @@ -43,7 +43,9 @@ UnigramDictionary::UnigramDictionary(const unsigned char *dict, int typedLetterM ROOT_POS(isLatestDictVersion ? DICTIONARY_HEADER_SIZE : 0), BYTES_IN_ONE_CHAR(MAX_PROXIMITY_CHARS * sizeof(*mInputCodes)), MAX_UMLAUT_SEARCH_DEPTH(DEFAULT_MAX_UMLAUT_SEARCH_DEPTH) { - if (DEBUG_DICT) LOGI("UnigramDictionary - constructor"); + if (DEBUG_DICT) { + LOGI("UnigramDictionary - constructor"); + } } UnigramDictionary::~UnigramDictionary() {} @@ -183,7 +185,9 @@ void UnigramDictionary::getWordSuggestions(const ProximityInfo *proximityInfo, // Suggestion with missing character if (SUGGEST_WORDS_WITH_MISSING_CHARACTER) { for (int i = 0; i < codesSize; ++i) { - if (DEBUG_DICT) LOGI("--- Suggest missing characters %d", i); + if (DEBUG_DICT) { + LOGI("--- Suggest missing characters %d", i); + } getSuggestionCandidates(i, -1, -1, NULL, 0, MAX_DEPTH); } } @@ -194,7 +198,9 @@ void UnigramDictionary::getWordSuggestions(const ProximityInfo *proximityInfo, if (SUGGEST_WORDS_WITH_EXCESSIVE_CHARACTER && mInputLength >= MIN_USER_TYPED_LENGTH_FOR_EXCESSIVE_CHARACTER_SUGGESTION) { for (int i = 0; i < codesSize; ++i) { - if (DEBUG_DICT) LOGI("--- Suggest excessive characters %d", i); + if (DEBUG_DICT) { + LOGI("--- Suggest excessive characters %d", i); + } getSuggestionCandidates(-1, i, -1, NULL, 0, MAX_DEPTH); } } @@ -205,7 +211,9 @@ void UnigramDictionary::getWordSuggestions(const ProximityInfo *proximityInfo, // Only suggest words that length is mInputLength if (SUGGEST_WORDS_WITH_TRANSPOSED_CHARACTERS) { for (int i = 0; i < codesSize; ++i) { - if (DEBUG_DICT) LOGI("--- Suggest transposed characters %d", i); + if (DEBUG_DICT) { + LOGI("--- Suggest transposed characters %d", i); + } getSuggestionCandidates(-1, -1, i, NULL, 0, mInputLength - 1); } } @@ -216,22 +224,27 @@ void UnigramDictionary::getWordSuggestions(const ProximityInfo *proximityInfo, if (SUGGEST_WORDS_WITH_MISSING_SPACE_CHARACTER && mInputLength >= MIN_USER_TYPED_LENGTH_FOR_MISSING_SPACE_SUGGESTION) { for (int i = 1; i < codesSize; ++i) { - if (DEBUG_DICT) LOGI("--- Suggest missing space characters %d", i); + if (DEBUG_DICT) { + LOGI("--- Suggest missing space characters %d", i); + } getMissingSpaceWords(mInputLength, i); } } PROF_END(5); PROF_START(6); - if (SUGGEST_WORDS_WITH_SPACE_PROXIMITY) { + if (SUGGEST_WORDS_WITH_SPACE_PROXIMITY && proximityInfo) { // The first and last "mistyped spaces" are taken care of by excessive character handling for (int i = 1; i < codesSize - 1; ++i) { - if (DEBUG_DICT) LOGI("--- Suggest words with proximity space %d", i); + if (DEBUG_DICT) { + LOGI("--- Suggest words with proximity space %d", i); + } const int x = xcoordinates[i]; const int y = ycoordinates[i]; - if (DEBUG_PROXIMITY_INFO) + if (DEBUG_PROXIMITY_INFO) { LOGI("Input[%d] x = %d, y = %d, has space proximity = %d", i, x, y, proximityInfo->hasSpaceProximity(x, y)); + } if (proximityInfo->hasSpaceProximity(x, y)) { getMistypedSpaceWords(mInputLength, i); } @@ -242,7 +255,9 @@ void UnigramDictionary::getWordSuggestions(const ProximityInfo *proximityInfo, void UnigramDictionary::initSuggestions(const int *codes, const int codesSize, unsigned short *outWords, int *frequencies) { - if (DEBUG_DICT) LOGI("initSuggest"); + if (DEBUG_DICT) { + LOGI("initSuggest"); + } mFrequencies = frequencies; mOutputChars = outWords; mInputCodes = codes; @@ -266,7 +281,9 @@ bool UnigramDictionary::addWord(unsigned short *word, int length, int frequency) LOGI("Found word = %s, freq = %d", s, frequency); } if (length > MAX_WORD_LENGTH) { - if (DEBUG_DICT) LOGI("Exceeded max word length."); + if (DEBUG_DICT) { + LOGI("Exceeded max word length."); + } return false; } @@ -283,7 +300,7 @@ bool UnigramDictionary::addWord(unsigned short *word, int length, int frequency) if (DEBUG_DICT) { char s[length + 1]; for (int i = 0; i <= length; i++) s[i] = word[i]; - LOGI("Added word = %s, freq = %d", s, frequency); + LOGI("Added word = %s, freq = %d, %d", s, frequency, S_INT_MAX); } memmove((char*) mFrequencies + (insertAt + 1) * sizeof(mFrequencies[0]), (char*) mFrequencies + insertAt * sizeof(mFrequencies[0]), @@ -297,7 +314,9 @@ bool UnigramDictionary::addWord(unsigned short *word, int length, int frequency) *dest++ = *word++; } *dest = 0; // NULL terminate - if (DEBUG_DICT) LOGI("Added word at %d", insertAt); + if (DEBUG_DICT) { + LOGI("Added word at %d", insertAt); + } return true; } return false; @@ -390,17 +409,105 @@ void UnigramDictionary::getSuggestionCandidates(const int skipPos, } } -inline static void multiplyRate(const int rate, int *freq) { - if (rate > 1000000) { - *freq = (*freq / 100) * rate; +static const int TWO_31ST_DIV_255 = S_INT_MAX / 255; +static inline int capped255MultForFullMatchAccentsOrCapitalizationDifference(const int num) { + return (num < TWO_31ST_DIV_255 ? 255 * num : S_INT_MAX); +} + +static const int TWO_31ST_DIV_2 = S_INT_MAX / 2; +inline static void multiplyIntCapped(const int multiplier, int *base) { + const int temp = *base; + if (temp != S_INT_MAX) { + // Branch if multiplier == 2 for the optimization + if (multiplier == 2) { + *base = TWO_31ST_DIV_2 >= temp ? temp << 1 : S_INT_MAX; + } else { + const int tempRetval = temp * multiplier; + *base = tempRetval >= temp ? tempRetval : S_INT_MAX; + } + } +} + +inline static int powerIntCapped(const int base, const int n) { + if (base == 2) { + return n < 31 ? 1 << n : S_INT_MAX; } else { - *freq = *freq * rate / 100; + int ret = base; + for (int i = 1; i < n; ++i) multiplyIntCapped(base, &ret); + return ret; + } +} + +inline static void multiplyRate(const int rate, int *freq) { + if (*freq != S_INT_MAX) { + if (*freq > 1000000) { + *freq /= 100; + multiplyIntCapped(rate, freq); + } else { + multiplyIntCapped(rate, freq); + *freq /= 100; + } + } +} + +inline static int calcFreqForSplitTwoWords( + const int typedLetterMultiplier, const int firstWordLength, const int secondWordLength, + const int firstFreq, const int secondFreq, const bool isSpaceProximity) { + if (firstWordLength == 0 || secondWordLength == 0) { + return 0; + } + const int firstDemotionRate = 100 - 100 / (firstWordLength + 1); + int tempFirstFreq = firstFreq; + multiplyRate(firstDemotionRate, &tempFirstFreq); + + const int secondDemotionRate = 100 - 100 / (secondWordLength + 1); + int tempSecondFreq = secondFreq; + multiplyRate(secondDemotionRate, &tempSecondFreq); + + const int totalLength = firstWordLength + secondWordLength; + + // Promote pairFreq with multiplying by 2, because the word length is the same as the typed + // length. + int totalFreq = tempFirstFreq + tempSecondFreq; + + // This is a workaround to try offsetting the not-enough-demotion which will be done in + // calcNormalizedScore in Utils.java. + // In calcNormalizedScore the score will be demoted by (1 - 1 / length) + // but we demoted only (1 - 1 / (length + 1)) so we will additionally adjust freq by + // (1 - 1 / length) / (1 - 1 / (length + 1)) = (1 - 1 / (length * length)) + const int normalizedScoreNotEnoughDemotionAdjustment = 100 - 100 / (totalLength * totalLength); + multiplyRate(normalizedScoreNotEnoughDemotionAdjustment, &totalFreq); + + // At this moment, totalFreq is calculated by the following formula: + // (firstFreq * (1 - 1 / (firstWordLength + 1)) + secondFreq * (1 - 1 / (secondWordLength + 1))) + // * (1 - 1 / totalLength) / (1 - 1 / (totalLength + 1)) + + multiplyIntCapped(powerIntCapped(typedLetterMultiplier, totalLength), &totalFreq); + + // This is another workaround to offset the demotion which will be done in + // calcNormalizedScore in Utils.java. + // In calcNormalizedScore the score will be demoted by (1 - 1 / length) so we have to promote + // the same amount because we already have adjusted the synthetic freq of this "missing or + // mistyped space" suggestion candidate above in this method. + const int normalizedScoreDemotionRateOffset = (100 + 100 / totalLength); + multiplyRate(normalizedScoreDemotionRateOffset, &totalFreq); + + if (isSpaceProximity) { + // A word pair with one space proximity correction + if (DEBUG_DICT) { + LOGI("Found a word pair with space proximity correction."); + } + multiplyIntCapped(typedLetterMultiplier, &totalFreq); + multiplyRate(WORDS_WITH_PROXIMITY_CHARACTER_DEMOTION_RATE, &totalFreq); } + + multiplyRate(WORDS_WITH_MISSING_SPACE_CHARACTER_DEMOTION_RATE, &totalFreq); + return totalFreq; } bool UnigramDictionary::getSplitTwoWordsSuggestion(const int inputLength, const int firstWordStartPos, const int firstWordLength, const int secondWordStartPos, - const int secondWordLength) { + const int secondWordLength, const bool isSpaceProximity) { if (inputLength >= MAX_WORD_LENGTH) return false; if (0 >= firstWordLength || 0 >= secondWordLength || firstWordStartPos >= secondWordStartPos || firstWordStartPos < 0 || secondWordStartPos + secondWordLength > inputLength) @@ -409,7 +516,9 @@ bool UnigramDictionary::getSplitTwoWordsSuggestion(const int inputLength, // Allocating variable length array on stack unsigned short word[newWordLength]; const int firstFreq = getBestWordFreq(firstWordStartPos, firstWordLength, mWord); - if (DEBUG_DICT) LOGI("First freq: %d", firstFreq); + if (DEBUG_DICT) { + LOGI("First freq: %d", firstFreq); + } if (firstFreq <= 0) return false; for (int i = 0; i < firstWordLength; ++i) { @@ -417,7 +526,9 @@ bool UnigramDictionary::getSplitTwoWordsSuggestion(const int inputLength, } const int secondFreq = getBestWordFreq(secondWordStartPos, secondWordLength, mWord); - if (DEBUG_DICT) LOGI("Second freq: %d", secondFreq); + if (DEBUG_DICT) { + LOGI("Second freq: %d", secondFreq); + } if (secondFreq <= 0) return false; word[firstWordLength] = SPACE; @@ -425,22 +536,25 @@ bool UnigramDictionary::getSplitTwoWordsSuggestion(const int inputLength, word[i] = mWord[i - firstWordLength - 1]; } - int pairFreq = ((firstFreq + secondFreq) / 2); - for (int i = 0; i < inputLength; ++i) pairFreq *= TYPED_LETTER_MULTIPLIER; - multiplyRate(WORDS_WITH_MISSING_SPACE_CHARACTER_DEMOTION_RATE, &pairFreq); + int pairFreq = calcFreqForSplitTwoWords(TYPED_LETTER_MULTIPLIER, firstWordLength, + secondWordLength, firstFreq, secondFreq, isSpaceProximity); + if (DEBUG_DICT) { + LOGI("Split two words: %d, %d, %d, %d, %d", firstFreq, secondFreq, pairFreq, inputLength, + TYPED_LETTER_MULTIPLIER); + } addWord(word, newWordLength, pairFreq); return true; } bool UnigramDictionary::getMissingSpaceWords(const int inputLength, const int missingSpacePos) { return getSplitTwoWordsSuggestion( - inputLength, 0, missingSpacePos, missingSpacePos, inputLength - missingSpacePos); + inputLength, 0, missingSpacePos, missingSpacePos, inputLength - missingSpacePos, false); } bool UnigramDictionary::getMistypedSpaceWords(const int inputLength, const int spaceProximityPos) { return getSplitTwoWordsSuggestion( inputLength, 0, spaceProximityPos, spaceProximityPos + 1, - inputLength - spaceProximityPos - 1); + inputLength - spaceProximityPos - 1, true); } // Keep this for comparing spec to new getWords @@ -485,19 +599,21 @@ void UnigramDictionary::getWordsRec(const int childrenCount, const int pos, cons } } -static const int TWO_31ST_DIV_255 = S_INT_MAX / 255; -static inline int capped255MultForFullMatchAccentsOrCapitalizationDifference(const int num) { - return (num < TWO_31ST_DIV_255 ? 255 * num : S_INT_MAX); -} inline int UnigramDictionary::calculateFinalFreq(const int inputIndex, const int depth, const int matchWeight, const int skipPos, const int excessivePos, const int transposedPos, const int freq, const bool sameLength) const { // TODO: Demote by edit distance int finalFreq = freq * matchWeight; if (skipPos >= 0) { - if (mInputLength >= 3) { - multiplyRate(WORDS_WITH_MISSING_CHARACTER_DEMOTION_RATE * - (mInputLength - 2) / (mInputLength - 1), &finalFreq); + if (mInputLength >= 2) { + const int demotionRate = WORDS_WITH_MISSING_CHARACTER_DEMOTION_RATE + * (10 * mInputLength - WORDS_WITH_MISSING_CHARACTER_DEMOTION_START_POS_10X) + / (10 * mInputLength + - WORDS_WITH_MISSING_CHARACTER_DEMOTION_START_POS_10X + 10); + if (DEBUG_DICT_FULL) { + LOGI("Demotion rate for missing character is %d.", demotionRate); + } + multiplyRate(demotionRate, &finalFreq); } else { finalFreq = 0; } @@ -511,17 +627,30 @@ inline int UnigramDictionary::calculateFinalFreq(const int inputIndex, const int } } int lengthFreq = TYPED_LETTER_MULTIPLIER; - for (int i = 0; i < depth; ++i) lengthFreq *= TYPED_LETTER_MULTIPLIER; + multiplyIntCapped(powerIntCapped(TYPED_LETTER_MULTIPLIER, depth), &lengthFreq); if (lengthFreq == matchWeight) { + // Full exact match if (depth > 1) { - if (DEBUG_DICT) LOGI("Found full matched word."); + if (DEBUG_DICT) { + LOGI("Found full matched word."); + } multiplyRate(FULL_MATCHED_WORDS_PROMOTION_RATE, &finalFreq); } if (sameLength && transposedPos < 0 && skipPos < 0 && excessivePos < 0) { finalFreq = capped255MultForFullMatchAccentsOrCapitalizationDifference(finalFreq); } + } else if (sameLength && transposedPos < 0 && skipPos < 0 && excessivePos < 0 && depth > 0) { + // A word with proximity corrections + if (DEBUG_DICT) { + LOGI("Found one proximity correction."); + } + multiplyIntCapped(TYPED_LETTER_MULTIPLIER, &finalFreq); + multiplyRate(WORDS_WITH_PROXIMITY_CHARACTER_DEMOTION_RATE, &finalFreq); + } + if (DEBUG_DICT) { + LOGI("calc: %d, %d", depth, sameLength); } - if (sameLength) finalFreq *= FULL_WORD_MULTIPLIER; + if (sameLength) multiplyIntCapped(FULL_WORD_MULTIPLIER, &finalFreq); return finalFreq; } @@ -674,7 +803,7 @@ inline bool UnigramDictionary::processCurrentNode(const int pos, const int depth // If inputIndex is greater than mInputLength, that means there is no // proximity chars. So, we don't need to check proximity. if (SAME_OR_ACCENTED_OR_CAPITALIZED_CHAR == matchedProximityCharId) { - matchWeight = matchWeight * TYPED_LETTER_MULTIPLIER; + multiplyIntCapped(TYPED_LETTER_MULTIPLIER, &matchWeight); } bool isSameAsUserTypedLength = mInputLength == inputIndex + 1 || (excessivePos == mInputLength - 1 && inputIndex == mInputLength - 2); @@ -768,7 +897,9 @@ inline bool UnigramDictionary::processCurrentNodeForExactMatch(const int firstCh *siblingPos = Dictionary::setDictionaryValues(DICT, IS_LATEST_DICT_VERSION, firstChildPos, &c, newChildPosition, newTerminal, newFreq); const unsigned int inputC = currentChars[0]; - if (DEBUG_DICT) assert(inputC <= U_SHORT_MAX); + if (DEBUG_DICT) { + assert(inputC <= U_SHORT_MAX); + } const unsigned short baseLowerC = toBaseLowerCase(c); const bool matched = (inputC == baseLowerC || inputC == c); const bool hasChild = *newChildPosition != 0; @@ -776,7 +907,9 @@ inline bool UnigramDictionary::processCurrentNodeForExactMatch(const int firstCh word[depth] = c; if (DEBUG_DICT && DEBUG_NODE) { LOGI("Node(%c, %c)<%d>, %d, %d", inputC, c, matched, hasChild, *newFreq); - if (*newTerminal) LOGI("Terminal %d", *newFreq); + if (*newTerminal) { + LOGI("Terminal %d", *newFreq); + } } if (hasChild) { *newCount = Dictionary::getCount(DICT, newChildPosition); diff --git a/native/src/unigram_dictionary.h b/native/src/unigram_dictionary.h index 3d3007ce0..dd1b89042 100644 --- a/native/src/unigram_dictionary.h +++ b/native/src/unigram_dictionary.h @@ -66,7 +66,7 @@ private: const int nextLettersSize); bool getSplitTwoWordsSuggestion(const int inputLength, const int firstWordStartPos, const int firstWordLength, - const int secondWordStartPos, const int secondWordLength); + const int secondWordStartPos, const int secondWordLength, const bool isSpaceProximity); bool getMissingSpaceWords(const int inputLength, const int missingSpacePos); bool getMistypedSpaceWords(const int inputLength, const int spaceProximityPos); // Keep getWordsOld for comparing performance between getWords and getWordsOld |