- separate dict (uses xml)
- retrieve bigrams that only starts with character typed and neighbor keys
- contacts bigram
- performance measure
bug: 2873133
Change-Id: If97c005b18c82f3fafef50009dd2dfd972b0ab8f
diff --git a/native/jni/com_android_inputmethod_latin_BinaryDictionary.cpp b/native/jni/com_android_inputmethod_latin_BinaryDictionary.cpp
index 4fe80da..bf7ec0d 100644
--- a/native/jni/com_android_inputmethod_latin_BinaryDictionary.cpp
+++ b/native/jni/com_android_inputmethod_latin_BinaryDictionary.cpp
@@ -67,8 +67,9 @@
int *nextLetters = nextLettersArray != NULL ? env->GetIntArrayElements(nextLettersArray, NULL)
: NULL;
- int count = dictionary->getSuggestions(inputCodes, arraySize, (unsigned short*) outputChars, frequencies,
- maxWordLength, maxWords, maxAlternatives, skipPos, nextLetters, nextLettersSize);
+ int count = dictionary->getSuggestions(inputCodes, arraySize, (unsigned short*) outputChars,
+ frequencies, maxWordLength, maxWords, maxAlternatives, skipPos, nextLetters,
+ nextLettersSize);
env->ReleaseIntArrayElements(frequencyArray, frequencies, 0);
env->ReleaseIntArrayElements(inputArray, inputCodes, JNI_ABORT);
@@ -81,20 +82,24 @@
}
static int latinime_BinaryDictionary_getBigrams
- (JNIEnv *env, jobject object, jint dict, jcharArray wordArray, jint wordLength,
- jcharArray outputArray, jintArray frequencyArray, jint maxWordLength, jint maxBigrams)
+ (JNIEnv *env, jobject object, jint dict, jcharArray prevWordArray, jint prevWordLength,
+ jintArray inputArray, jint inputArraySize, jcharArray outputArray,
+ jintArray frequencyArray, jint maxWordLength, jint maxBigrams, jint maxAlternatives)
{
Dictionary *dictionary = (Dictionary*) dict;
if (dictionary == NULL) return 0;
- jchar *word = env->GetCharArrayElements(wordArray, NULL);
+ jchar *prevWord = env->GetCharArrayElements(prevWordArray, NULL);
+ int *inputCodes = env->GetIntArrayElements(inputArray, NULL);
jchar *outputChars = env->GetCharArrayElements(outputArray, NULL);
int *frequencies = env->GetIntArrayElements(frequencyArray, NULL);
- int count = dictionary->getBigrams((unsigned short*) word, wordLength,
- (unsigned short*) outputChars, frequencies, maxWordLength, maxBigrams);
+ int count = dictionary->getBigrams((unsigned short*) prevWord, prevWordLength, inputCodes,
+ inputArraySize, (unsigned short*) outputChars, frequencies, maxWordLength, maxBigrams,
+ maxAlternatives);
- env->ReleaseCharArrayElements(wordArray, word, JNI_ABORT);
+ env->ReleaseCharArrayElements(prevWordArray, prevWord, JNI_ABORT);
+ env->ReleaseIntArrayElements(inputArray, inputCodes, JNI_ABORT);
env->ReleaseCharArrayElements(outputArray, outputChars, 0);
env->ReleaseIntArrayElements(frequencyArray, frequencies, 0);
@@ -130,7 +135,7 @@
{"closeNative", "(I)V", (void*)latinime_BinaryDictionary_close},
{"getSuggestionsNative", "(I[II[C[IIIII[II)I", (void*)latinime_BinaryDictionary_getSuggestions},
{"isValidWordNative", "(I[CI)Z", (void*)latinime_BinaryDictionary_isValidWord},
- {"getBigramsNative", "(I[CI[C[III)I", (void*)latinime_BinaryDictionary_getBigrams}
+ {"getBigramsNative", "(I[CI[II[C[IIII)I", (void*)latinime_BinaryDictionary_getBigrams}
};
static int registerNativeMethods(JNIEnv* env, const char* className,
diff --git a/native/src/dictionary.cpp b/native/src/dictionary.cpp
index a1a632f..1a39f585b4 100644
--- a/native/src/dictionary.cpp
+++ b/native/src/dictionary.cpp
@@ -387,13 +387,17 @@
int
-Dictionary::getBigrams(unsigned short *prevWord, int prevWordLength, unsigned short *bigramChars,
- int *bigramFreq, int maxWordLength, int maxBigrams)
+Dictionary::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;
mMaxWordLength = maxWordLength;
mMaxBigrams = maxBigrams;
+ mMaxAlternatives = maxAlternatives;
if (mBigram == 1 && checkIfDictVersionIsLatest()) {
int pos = isValidWordRec(DICTIONARY_HEADER_SIZE, prevWord, 0, prevWordLength);
@@ -406,7 +410,7 @@
int bigramExist = (mDict[pos] & FLAG_BIGRAM_READ);
if (bigramExist > 0) {
int nextBigramExist = 1;
- while (nextBigramExist > 0) {
+ while (nextBigramExist > 0 && bigramCount < maxBigrams) {
int bigramAddress = getBigramAddress(&pos, true);
int frequency = (FLAG_BIGRAM_FREQ & mDict[pos]);
// search for all bigrams and store them
@@ -521,8 +525,27 @@
break;
}
}
+ if (checkFirstCharacter(word)) {
+ addWordBigram(word, depth, frequency);
+ }
+}
- addWordBigram(word, depth, frequency);
+bool
+Dictionary::checkFirstCharacter(unsigned short *word)
+{
+ // Checks whether this word starts with same character or neighboring characters of
+ // what user typed.
+
+ int *inputCodes = mInputCodes;
+ int maxAlt = mMaxAlternatives;
+ while (maxAlt > 0) {
+ if ((unsigned int) *inputCodes == (unsigned int) *word) {
+ return true;
+ }
+ inputCodes++;
+ maxAlt--;
+ }
+ return false;
}
bool
diff --git a/native/src/dictionary.h b/native/src/dictionary.h
index 2c57429..d13496e 100644
--- a/native/src/dictionary.h
+++ b/native/src/dictionary.h
@@ -39,8 +39,9 @@
int getSuggestions(int *codes, int codesSize, unsigned short *outWords, int *frequencies,
int maxWordLength, int maxWords, int maxAlternatives, int skipPos,
int *nextLetters, int nextLettersSize);
- int getBigrams(unsigned short *word, int length, unsigned short *outWords, int *frequencies,
- int maxWordLength, int maxBigrams);
+ int getBigrams(unsigned short *word, int length, int *codes, int codesSize,
+ unsigned short *outWords, int *frequencies, int maxWordLength, int maxBigrams,
+ int maxAlternatives);
bool isValidWord(unsigned short *word, int length);
void setAsset(void *asset) { mAsset = asset; }
void *getAsset() { return mAsset; }
@@ -64,6 +65,7 @@
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);