diff --git a/native/src/bigram_dictionary.cpp b/native/src/bigram_dictionary.cpp
index f0d5f8c9afda08636beedf76befe3ed8c0cf69ca..095b80531fd88e5af539a08675d96ba3224f2621 100644
--- a/native/src/bigram_dictionary.cpp
+++ b/native/src/bigram_dictionary.cpp
@@ -15,19 +15,240 @@
** limitations under the License.
*/
+#include <string.h>
+
#define LOG_TAG "LatinIME: bigram_dictionary.cpp"
#include "bigram_dictionary.h"
+#include "dictionary.h"
namespace latinime {
-BigramDictionary::BigramDictionary(void *dict, int typedLetterMultiplier, int fullWordMultiplier,
- int maxWordLength, int maxWords, int maxAlternatives, Dictionary *parentDictionary)
-{
+BigramDictionary::BigramDictionary(const unsigned char *dict, int maxWordLength,
+ int maxAlternatives, const bool isLatestDictVersion, const bool hasBigram,
+ Dictionary *parentDictionary)
+ : DICT(dict), MAX_WORD_LENGTH(maxWordLength),
+ MAX_ALTERNATIVES(maxAlternatives), IS_LATEST_DICT_VERSION(isLatestDictVersion),
+ HAS_BIGRAM(hasBigram), mParentDictionary(parentDictionary) {
+ LOGI("BigramDictionary - constructor");
+ LOGI("Has Bigram : %d \n", hasBigram);
+}
+
+BigramDictionary::~BigramDictionary() {
+}
+
+bool BigramDictionary::addWordBigram(unsigned short *word, int length, int frequency) {
+ word[length] = 0;
+ if (DEBUG_DICT) {
+ char s[length + 1];
+ for (int i = 0; i <= length; i++) s[i] = word[i];
+ LOGI("Bigram: Found word = %s, freq = %d : \n", s, frequency);
+ }
+
+ // Find the right insertion point
+ int insertAt = 0;
+ while (insertAt < mMaxBigrams) {
+ if (frequency > mBigramFreq[insertAt] || (mBigramFreq[insertAt] == frequency
+ && length < Dictionary::wideStrLen(mBigramChars + insertAt * MAX_WORD_LENGTH))) {
+ break;
+ }
+ insertAt++;
+ }
+ LOGI("Bigram: InsertAt -> %d maxBigrams: %d\n", insertAt, mMaxBigrams);
+ if (insertAt < mMaxBigrams) {
+ memmove((char*) mBigramFreq + (insertAt + 1) * sizeof(mBigramFreq[0]),
+ (char*) mBigramFreq + insertAt * sizeof(mBigramFreq[0]),
+ (mMaxBigrams - insertAt - 1) * sizeof(mBigramFreq[0]));
+ mBigramFreq[insertAt] = frequency;
+ memmove((char*) mBigramChars + (insertAt + 1) * MAX_WORD_LENGTH * sizeof(short),
+ (char*) mBigramChars + (insertAt ) * MAX_WORD_LENGTH * sizeof(short),
+ (mMaxBigrams - insertAt - 1) * sizeof(short) * MAX_WORD_LENGTH);
+ unsigned short *dest = mBigramChars + (insertAt ) * MAX_WORD_LENGTH;
+ while (length--) {
+ *dest++ = *word++;
+ }
+ *dest = 0; // NULL terminate
+ if (DEBUG_DICT) LOGI("Bigram: Added word at %d\n", insertAt);
+ return true;
+ }
+ return false;
+}
+
+int BigramDictionary::getBigramAddress(int *pos, bool advance) {
+ int address = 0;
+
+ address += (DICT[*pos] & 0x3F) << 16;
+ address += (DICT[*pos + 1] & 0xFF) << 8;
+ address += (DICT[*pos + 2] & 0xFF);
+
+ if (advance) {
+ *pos += 3;
+ }
+
+ return address;
+}
+
+int BigramDictionary::getBigramFreq(int *pos) {
+ int freq = DICT[(*pos)++] & FLAG_BIGRAM_FREQ;
+
+ return freq;
}
-BigramDictionary::~BigramDictionary()
-{
+
+int BigramDictionary::getBigrams(unsigned short *prevWord, int prevWordLength, int *codes,
+ int codesSize, unsigned short *bigramChars, int *bigramFreq, int maxWordLength,
+ int maxBigrams, int maxAlternatives) {
+ mBigramFreq = bigramFreq;
+ mBigramChars = bigramChars;
+ mInputCodes = codes;
+ mInputLength = codesSize;
+ mMaxBigrams = maxBigrams;
+
+ if (HAS_BIGRAM && IS_LATEST_DICT_VERSION) {
+ int pos = mParentDictionary->isValidWordRec(
+ DICTIONARY_HEADER_SIZE, prevWord, 0, prevWordLength);
+ LOGI("Pos -> %d\n", pos);
+ if (pos < 0) {
+ return 0;
+ }
+
+ int bigramCount = 0;
+ int bigramExist = (DICT[pos] & FLAG_BIGRAM_READ);
+ if (bigramExist > 0) {
+ int nextBigramExist = 1;
+ while (nextBigramExist > 0 && bigramCount < maxBigrams) {
+ int bigramAddress = getBigramAddress(&pos, true);
+ int frequency = (FLAG_BIGRAM_FREQ & DICT[pos]);
+ // search for all bigrams and store them
+ searchForTerminalNode(bigramAddress, frequency);
+ nextBigramExist = (DICT[pos++] & FLAG_BIGRAM_CONTINUED);
+ bigramCount++;
+ }
+ }
+
+ return bigramCount;
+ }
+ return 0;
+}
+
+void BigramDictionary::searchForTerminalNode(int addressLookingFor, int frequency) {
+ // track word with such address and store it in an array
+ unsigned short word[MAX_WORD_LENGTH];
+
+ int pos;
+ int followDownBranchAddress = DICTIONARY_HEADER_SIZE;
+ bool found = false;
+ char followingChar = ' ';
+ int depth = -1;
+
+ while(!found) {
+ bool followDownAddressSearchStop = false;
+ bool firstAddress = true;
+ bool haveToSearchAll = true;
+
+ if (depth >= 0) {
+ word[depth] = (unsigned short) followingChar;
+ }
+ pos = followDownBranchAddress; // pos start at count
+ int count = DICT[pos] & 0xFF;
+ LOGI("count - %d\n",count);
+ pos++;
+ for (int i = 0; i < count; i++) {
+ // pos at data
+ pos++;
+ // pos now at flag
+ if (!getFirstBitOfByte(&pos)) { // non-terminal
+ if (!followDownAddressSearchStop) {
+ int addr = getBigramAddress(&pos, false);
+ if (addr > addressLookingFor) {
+ followDownAddressSearchStop = true;
+ if (firstAddress) {
+ firstAddress = false;
+ haveToSearchAll = true;
+ } else if (!haveToSearchAll) {
+ break;
+ }
+ } else {
+ followDownBranchAddress = addr;
+ followingChar = (char)(0xFF & DICT[pos-1]);
+ if (firstAddress) {
+ firstAddress = false;
+ haveToSearchAll = false;
+ }
+ }
+ }
+ pos += 3;
+ } else if (getFirstBitOfByte(&pos)) { // terminal
+ if (addressLookingFor == (pos-1)) { // found !!
+ depth++;
+ word[depth] = (0xFF & DICT[pos-1]);
+ found = true;
+ break;
+ }
+ if (getSecondBitOfByte(&pos)) { // address + freq (4 byte)
+ if (!followDownAddressSearchStop) {
+ int addr = getBigramAddress(&pos, false);
+ if (addr > addressLookingFor) {
+ followDownAddressSearchStop = true;
+ if (firstAddress) {
+ firstAddress = false;
+ haveToSearchAll = true;
+ } else if (!haveToSearchAll) {
+ break;
+ }
+ } else {
+ followDownBranchAddress = addr;
+ followingChar = (char)(0xFF & DICT[pos-1]);
+ if (firstAddress) {
+ firstAddress = false;
+ haveToSearchAll = true;
+ }
+ }
+ }
+ pos += 4;
+ } else { // freq only (2 byte)
+ pos += 2;
+ }
+
+ // skipping bigram
+ int bigramExist = (DICT[pos] & FLAG_BIGRAM_READ);
+ if (bigramExist > 0) {
+ int nextBigramExist = 1;
+ while (nextBigramExist > 0) {
+ pos += 3;
+ nextBigramExist = (DICT[pos++] & FLAG_BIGRAM_CONTINUED);
+ }
+ } else {
+ pos++;
+ }
+ }
+ }
+ depth++;
+ if (followDownBranchAddress == 0) {
+ LOGI("ERROR!!! Cannot find bigram!!");
+ break;
+ }
+ }
+ if (checkFirstCharacter(word)) {
+ addWordBigram(word, depth, frequency);
+ }
}
+
+bool BigramDictionary::checkFirstCharacter(unsigned short *word) {
+ // Checks whether this word starts with same character or neighboring characters of
+ // what user typed.
+
+ int *inputCodes = mInputCodes;
+ int maxAlt = MAX_ALTERNATIVES;
+ while (maxAlt > 0) {
+ if ((unsigned int) *inputCodes == (unsigned int) *word) {
+ return true;
+ }
+ inputCodes++;
+ maxAlt--;
+ }
+ return false;
+}
+
// TODO: Move functions related to bigram to here
} // namespace latinime
diff --git a/native/src/bigram_dictionary.h b/native/src/bigram_dictionary.h
index 98146a0e7e44a024276e5851cca27ab46e989d06..d658b93e620068b2c16a036b20efb7d7b1ee946e 100644
--- a/native/src/bigram_dictionary.h
+++ b/native/src/bigram_dictionary.h
@@ -22,10 +22,33 @@ namespace latinime {
class Dictionary;
class BigramDictionary {
public:
- BigramDictionary(void *dict, int typedLetterMultipler, int fullWordMultiplier, int maxWordLength,
- int maxWords, int maxAlternatives, Dictionary *parentDictionary);
+ BigramDictionary(const unsigned char *dict, int maxWordLength, int maxAlternatives,
+ const bool isLatestDictVersion, const bool hasBigram, Dictionary *parentDictionary);
+ int getBigrams(unsigned short *word, int length, int *codes, int codesSize,
+ unsigned short *outWords, int *frequencies, int maxWordLength, int maxBigrams,
+ int maxAlternatives);
~BigramDictionary();
private:
+ bool addWordBigram(unsigned short *word, int length, int frequency);
+ int getBigramAddress(int *pos, bool advance);
+ int getBigramFreq(int *pos);
+ void searchForTerminalNode(int addressLookingFor, int frequency);
+ bool getFirstBitOfByte(int *pos) { return (DICT[*pos] & 0x80) > 0; }
+ bool getSecondBitOfByte(int *pos) { return (DICT[*pos] & 0x40) > 0; }
+ bool checkFirstCharacter(unsigned short *word);
+
+ const unsigned char *DICT;
+ const int MAX_WORD_LENGTH;
+ const int MAX_ALTERNATIVES;
+ const bool IS_LATEST_DICT_VERSION;
+ const bool HAS_BIGRAM;
+
+ Dictionary *mParentDictionary;
+ int *mBigramFreq;
+ int mMaxBigrams;
+ unsigned short *mBigramChars;
+ int *mInputCodes;
+ int mInputLength;
};
// ----------------------------------------------------------------------------
}; // namespace latinime
diff --git a/native/src/dictionary.cpp b/native/src/dictionary.cpp
index cf050fd30ba10782595343a322cd620b0f3dc44f..05692f7ef8e969eb19f7d4f05d3cba74b22d82cf 100644
--- a/native/src/dictionary.cpp
+++ b/native/src/dictionary.cpp
@@ -31,10 +31,9 @@ Dictionary::Dictionary(void *dict, int typedLetterMultiplier, int fullWordMultip
{
LOGI("IN NATIVE SUGGEST Version: %d \n", (DICT[0] & 0xFF));
mUnigramDictionary = new UnigramDictionary(DICT, typedLetterMultiplier, fullWordMultiplier,
- maxWordLength, maxWords, maxAlternatives, IS_LATEST_DICT_VERSION,
- hasBigram(), this);
- mBigramDictionary = new BigramDictionary(dict, typedLetterMultiplier, fullWordMultiplier,
- maxWordLength, maxWords, maxAlternatives, this);
+ maxWordLength, maxWords, maxAlternatives, IS_LATEST_DICT_VERSION);
+ mBigramDictionary = new BigramDictionary(DICT, maxWordLength, maxAlternatives,
+ IS_LATEST_DICT_VERSION, hasBigram(), this);
}
Dictionary::~Dictionary()
diff --git a/native/src/dictionary.h b/native/src/dictionary.h
index 61f7cf07499683a0a079fce180ec63c35a3734d5..1cd517cfb9b4e503e2f88cb46ccf9c20160e3fdf 100644
--- a/native/src/dictionary.h
+++ b/native/src/dictionary.h
@@ -37,7 +37,7 @@ public:
int getBigrams(unsigned short *word, int length, int *codes, int codesSize,
unsigned short *outWords, int *frequencies, int maxWordLength, int maxBigrams,
int maxAlternatives) {
- return mUnigramDictionary->getBigrams(word, length, codes, codesSize, outWords, frequencies,
+ return mBigramDictionary->getBigrams(word, length, codes, codesSize, outWords, frequencies,
maxWordLength, maxBigrams, maxAlternatives);
}
bool isValidWord(unsigned short *word, int length);
@@ -53,6 +53,7 @@ public:
static bool getTerminal(const unsigned char *dict, int *pos);
static int getAddress(const unsigned char *dict, int *pos);
static int getFreq(const unsigned char *dict, const bool isLatestDictVersion, int *pos);
+ static int wideStrLen(unsigned short *str);
private:
bool hasBigram();
@@ -117,5 +118,14 @@ inline int Dictionary::getFreq(const unsigned char *dict,
return freq;
}
+
+inline int Dictionary::wideStrLen(unsigned short *str) {
+ if (!str) return 0;
+ unsigned short *end = str;
+ while (*end)
+ end++;
+ return end - str;
+}
+
}; // namespace latinime
#endif // LATINIME_DICTIONARY_H
diff --git a/native/src/unigram_dictionary.cpp b/native/src/unigram_dictionary.cpp
index 5a48a97a8a48f3e904878db1443c20bbc19fe124..d0c903e81ee35e25409b0847452b044ed7e555ab 100644
--- a/native/src/unigram_dictionary.cpp
+++ b/native/src/unigram_dictionary.cpp
@@ -31,23 +31,17 @@ namespace latinime {
UnigramDictionary::UnigramDictionary(const unsigned char *dict, int typedLetterMultiplier,
int fullWordMultiplier, int maxWordLength, int maxWords, int maxAlternatives,
- const bool isLatestDictVersion, const bool hasBigram, Dictionary *parentDictionary)
+ const bool isLatestDictVersion)
: DICT(dict), MAX_WORD_LENGTH(maxWordLength),MAX_WORDS(maxWords),
MAX_ALTERNATIVES(maxAlternatives), IS_LATEST_DICT_VERSION(isLatestDictVersion),
- HAS_BIGRAM(hasBigram), mParentDictionary(parentDictionary)
-{
+ TYPED_LETTER_MULTIPLIER(typedLetterMultiplier), FULL_WORD_MULTIPLIER(fullWordMultiplier) {
LOGI("UnigramDictionary - constructor");
- LOGI("Has Bigram : %d \n", hasBigram);
- mTypedLetterMultiplier = typedLetterMultiplier;
- mFullWordMultiplier = fullWordMultiplier;
}
-UnigramDictionary::~UnigramDictionary()
-{
-}
+UnigramDictionary::~UnigramDictionary() {}
-int UnigramDictionary::getSuggestions(int *codes, int codesSize, unsigned short *outWords, int *frequencies,
- int *nextLetters, int nextLettersSize)
+int UnigramDictionary::getSuggestions(int *codes, int codesSize, unsigned short *outWords,
+ int *frequencies, int *nextLetters, int nextLettersSize)
{
initSuggestions(codes, codesSize, outWords, frequencies);
@@ -114,16 +108,6 @@ void UnigramDictionary::registerNextLetter(unsigned short c, int *nextLetters, i
}
}
-int
-UnigramDictionary::wideStrLen(unsigned short *str)
-{
- if (!str) return 0;
- unsigned short *end = str;
- while (*end)
- end++;
- return end - str;
-}
-
bool
UnigramDictionary::addWord(unsigned short *word, int length, int frequency)
{
@@ -139,7 +123,7 @@ UnigramDictionary::addWord(unsigned short *word, int length, int frequency)
while (insertAt < MAX_WORDS) {
if (frequency > mFrequencies[insertAt]
|| (mFrequencies[insertAt] == frequency
- && length < wideStrLen(mOutputChars + insertAt * MAX_WORD_LENGTH))) {
+ && length < Dictionary::wideStrLen(mOutputChars + insertAt * MAX_WORD_LENGTH))) {
break;
}
insertAt++;
@@ -163,46 +147,6 @@ UnigramDictionary::addWord(unsigned short *word, int length, int frequency)
return false;
}
-bool
-UnigramDictionary::addWordBigram(unsigned short *word, int length, int frequency)
-{
- word[length] = 0;
- if (DEBUG_DICT) {
- char s[length + 1];
- for (int i = 0; i <= length; i++) s[i] = word[i];
- LOGI("Bigram: Found word = %s, freq = %d : \n", s, frequency);
- }
-
- // Find the right insertion point
- int insertAt = 0;
- while (insertAt < mMaxBigrams) {
- if (frequency > mBigramFreq[insertAt]
- || (mBigramFreq[insertAt] == frequency
- && length < wideStrLen(mBigramChars + insertAt * MAX_WORD_LENGTH))) {
- break;
- }
- insertAt++;
- }
- LOGI("Bigram: InsertAt -> %d maxBigrams: %d\n", insertAt, mMaxBigrams);
- if (insertAt < mMaxBigrams) {
- memmove((char*) mBigramFreq + (insertAt + 1) * sizeof(mBigramFreq[0]),
- (char*) mBigramFreq + insertAt * sizeof(mBigramFreq[0]),
- (mMaxBigrams - insertAt - 1) * sizeof(mBigramFreq[0]));
- mBigramFreq[insertAt] = frequency;
- memmove((char*) mBigramChars + (insertAt + 1) * MAX_WORD_LENGTH * sizeof(short),
- (char*) mBigramChars + (insertAt ) * MAX_WORD_LENGTH * sizeof(short),
- (mMaxBigrams - insertAt - 1) * sizeof(short) * MAX_WORD_LENGTH);
- unsigned short *dest = mBigramChars + (insertAt ) * MAX_WORD_LENGTH;
- while (length--) {
- *dest++ = *word++;
- }
- *dest = 0; // NULL terminate
- if (DEBUG_DICT) LOGI("Bigram: Added word at %d\n", insertAt);
- return true;
- }
- return false;
-}
-
unsigned short
UnigramDictionary::toLowerCase(unsigned short c) {
if (c < sizeof(BASE_CHARS) / sizeof(BASE_CHARS[0])) {
@@ -236,8 +180,8 @@ UnigramDictionary::sameAsTyped(unsigned short *word, int length)
static char QUOTE = '\'';
void
-UnigramDictionary::getWordsRec(int pos, int depth, int maxDepth, bool completion, int snr, int inputIndex,
- int diffs, int skipPos, int *nextLetters, int nextLettersSize)
+UnigramDictionary::getWordsRec(int pos, int depth, int maxDepth, bool completion, int snr,
+ int inputIndex, int diffs, int skipPos, int *nextLetters, int nextLettersSize)
{
// Optimization: Prune out words that are too long compared to how much was typed.
if (depth > maxDepth) {
@@ -290,14 +234,14 @@ UnigramDictionary::getWordsRec(int pos, int depth, int maxDepth, bool completion
int j = 0;
while (currentChars[j] > 0) {
if (currentChars[j] == lowerC || currentChars[j] == c) {
- int addedWeight = j == 0 ? mTypedLetterMultiplier : 1;
+ int addedWeight = j == 0 ? TYPED_LETTER_MULTIPLIER : 1;
mWord[depth] = c;
if (mInputLength == inputIndex + 1) {
if (terminal) {
if (//INCLUDE_TYPED_WORD_IF_VALID ||
!sameAsTyped(mWord, depth + 1)) {
int finalFreq = freq * snr * addedWeight;
- if (skipPos < 0) finalFreq *= mFullWordMultiplier;
+ if (skipPos < 0) finalFreq *= FULL_WORD_MULTIPLIER;
addWord(mWord, depth + 1, finalFreq);
}
}
@@ -319,190 +263,4 @@ UnigramDictionary::getWordsRec(int pos, int depth, int maxDepth, bool completion
}
}
-int
-UnigramDictionary::getBigramAddress(int *pos, bool advance)
-{
- int address = 0;
-
- address += (DICT[*pos] & 0x3F) << 16;
- address += (DICT[*pos + 1] & 0xFF) << 8;
- address += (DICT[*pos + 2] & 0xFF);
-
- if (advance) {
- *pos += 3;
- }
-
- return address;
-}
-
-int
-UnigramDictionary::getBigramFreq(int *pos)
-{
- int freq = DICT[(*pos)++] & FLAG_BIGRAM_FREQ;
-
- return freq;
-}
-
-
-int
-UnigramDictionary::getBigrams(unsigned short *prevWord, int prevWordLength, int *codes, int codesSize,
- unsigned short *bigramChars, int *bigramFreq, int maxWordLength, int maxBigrams,
- int maxAlternatives)
-{
- mBigramFreq = bigramFreq;
- mBigramChars = bigramChars;
- mInputCodes = codes;
- mInputLength = codesSize;
- mMaxBigrams = maxBigrams;
-
- if (HAS_BIGRAM && IS_LATEST_DICT_VERSION) {
- int pos = mParentDictionary->isValidWordRec(
- DICTIONARY_HEADER_SIZE, prevWord, 0, prevWordLength);
- LOGI("Pos -> %d\n", pos);
- if (pos < 0) {
- return 0;
- }
-
- int bigramCount = 0;
- int bigramExist = (DICT[pos] & FLAG_BIGRAM_READ);
- if (bigramExist > 0) {
- int nextBigramExist = 1;
- while (nextBigramExist > 0 && bigramCount < maxBigrams) {
- int bigramAddress = getBigramAddress(&pos, true);
- int frequency = (FLAG_BIGRAM_FREQ & DICT[pos]);
- // search for all bigrams and store them
- searchForTerminalNode(bigramAddress, frequency);
- nextBigramExist = (DICT[pos++] & FLAG_BIGRAM_CONTINUED);
- bigramCount++;
- }
- }
-
- return bigramCount;
- }
- return 0;
-}
-
-void
-UnigramDictionary::searchForTerminalNode(int addressLookingFor, int frequency)
-{
- // track word with such address and store it in an array
- unsigned short word[MAX_WORD_LENGTH];
-
- int pos;
- int followDownBranchAddress = DICTIONARY_HEADER_SIZE;
- bool found = false;
- char followingChar = ' ';
- int depth = -1;
-
- while(!found) {
- bool followDownAddressSearchStop = false;
- bool firstAddress = true;
- bool haveToSearchAll = true;
-
- if (depth >= 0) {
- word[depth] = (unsigned short) followingChar;
- }
- pos = followDownBranchAddress; // pos start at count
- int count = DICT[pos] & 0xFF;
- LOGI("count - %d\n",count);
- pos++;
- for (int i = 0; i < count; i++) {
- // pos at data
- pos++;
- // pos now at flag
- if (!getFirstBitOfByte(&pos)) { // non-terminal
- if (!followDownAddressSearchStop) {
- int addr = getBigramAddress(&pos, false);
- if (addr > addressLookingFor) {
- followDownAddressSearchStop = true;
- if (firstAddress) {
- firstAddress = false;
- haveToSearchAll = true;
- } else if (!haveToSearchAll) {
- break;
- }
- } else {
- followDownBranchAddress = addr;
- followingChar = (char)(0xFF & DICT[pos-1]);
- if (firstAddress) {
- firstAddress = false;
- haveToSearchAll = false;
- }
- }
- }
- pos += 3;
- } else if (getFirstBitOfByte(&pos)) { // terminal
- if (addressLookingFor == (pos-1)) { // found !!
- depth++;
- word[depth] = (0xFF & DICT[pos-1]);
- found = true;
- break;
- }
- if (getSecondBitOfByte(&pos)) { // address + freq (4 byte)
- if (!followDownAddressSearchStop) {
- int addr = getBigramAddress(&pos, false);
- if (addr > addressLookingFor) {
- followDownAddressSearchStop = true;
- if (firstAddress) {
- firstAddress = false;
- haveToSearchAll = true;
- } else if (!haveToSearchAll) {
- break;
- }
- } else {
- followDownBranchAddress = addr;
- followingChar = (char)(0xFF & DICT[pos-1]);
- if (firstAddress) {
- firstAddress = false;
- haveToSearchAll = true;
- }
- }
- }
- pos += 4;
- } else { // freq only (2 byte)
- pos += 2;
- }
-
- // skipping bigram
- int bigramExist = (DICT[pos] & FLAG_BIGRAM_READ);
- if (bigramExist > 0) {
- int nextBigramExist = 1;
- while (nextBigramExist > 0) {
- pos += 3;
- nextBigramExist = (DICT[pos++] & FLAG_BIGRAM_CONTINUED);
- }
- } else {
- pos++;
- }
- }
- }
- depth++;
- if (followDownBranchAddress == 0) {
- LOGI("ERROR!!! Cannot find bigram!!");
- break;
- }
- }
- if (checkFirstCharacter(word)) {
- addWordBigram(word, depth, frequency);
- }
-}
-
-bool
-UnigramDictionary::checkFirstCharacter(unsigned short *word)
-{
- // Checks whether this word starts with same character or neighboring characters of
- // what user typed.
-
- int *inputCodes = mInputCodes;
- int maxAlt = MAX_ALTERNATIVES;
- while (maxAlt > 0) {
- if ((unsigned int) *inputCodes == (unsigned int) *word) {
- return true;
- }
- inputCodes++;
- maxAlt--;
- }
- return false;
-}
-
} // namespace latinime
diff --git a/native/src/unigram_dictionary.h b/native/src/unigram_dictionary.h
index 4cb1abe9b74bbf32644cabc5eab9cd55de86bc07..557e54cb7be81e96cd82607902528ca57a1e0bd9 100644
--- a/native/src/unigram_dictionary.h
+++ b/native/src/unigram_dictionary.h
@@ -24,63 +24,42 @@ namespace latinime {
class UnigramDictionary {
public:
UnigramDictionary(const unsigned char *dict, int typedLetterMultipler, int fullWordMultiplier,
- int maxWordLength, int maxWords, int maxAlternatives, const bool isLatestDictVersion,
- const bool hasBigram, Dictionary *parentDictionary);
+ int maxWordLength, int maxWords, int maxAlternatives, const bool isLatestDictVersion);
int getSuggestions(int *codes, int codesSize, unsigned short *outWords, int *frequencies,
int *nextLetters, int nextLettersSize);
- int getBigrams(unsigned short *word, int length, int *codes, int codesSize,
- unsigned short *outWords, int *frequencies, int maxWordLength, int maxBigrams,
- int maxAlternatives);
~UnigramDictionary();
private:
void initSuggestions(int *codes, int codesSize, unsigned short *outWords, int *frequencies);
- int getSuggestionCandidates(int inputLength, int skipPos, int *nextLetters, int nextLettersSize);
+ int getSuggestionCandidates(int inputLength, int skipPos, int *nextLetters,
+ int nextLettersSize);
void getVersionNumber();
bool checkIfDictVersionIsLatest();
int getAddress(int *pos);
- int getBigramAddress(int *pos, bool advance);
int getFreq(int *pos);
- int getBigramFreq(int *pos);
- void searchForTerminalNode(int address, int frequency);
-
- bool getFirstBitOfByte(int *pos) { return (DICT[*pos] & 0x80) > 0; }
- bool getSecondBitOfByte(int *pos) { return (DICT[*pos] & 0x40) > 0; }
- bool getTerminal(int *pos) { return (DICT[*pos] & FLAG_TERMINAL_MASK) > 0; }
int wideStrLen(unsigned short *str);
bool sameAsTyped(unsigned short *word, int length);
- bool checkFirstCharacter(unsigned short *word);
bool addWord(unsigned short *word, int length, int frequency);
- bool addWordBigram(unsigned short *word, int length, int frequency);
unsigned short toLowerCase(unsigned short c);
void getWordsRec(int pos, int depth, int maxDepth, bool completion, int frequency,
int inputIndex, int diffs, int skipPos, int *nextLetters, int nextLettersSize);
void registerNextLetter(unsigned short c, int *nextLetters, int nextLettersSize);
const unsigned char *DICT;
-
const int MAX_WORDS;
const int MAX_WORD_LENGTH;
const int MAX_ALTERNATIVES;
const bool IS_LATEST_DICT_VERSION;
- const bool HAS_BIGRAM;
+ const int TYPED_LETTER_MULTIPLIER;
+ const int FULL_WORD_MULTIPLIER;
- Dictionary *mParentDictionary;
int *mFrequencies;
- int *mBigramFreq;
- int mMaxBigrams;
unsigned short *mOutputChars;
- unsigned short *mBigramChars;
int *mInputCodes;
int mInputLength;
unsigned short mWord[128];
int mMaxEditDistance;
-
- int mFullWordMultiplier;
- int mTypedLetterMultiplier;
- int mVersion;
- int mBigram;
};
// ----------------------------------------------------------------------------