Merge "Return whether the dynamic dict operation was success."
This commit is contained in:
commit
3faf4fc5e7
13 changed files with 151 additions and 91 deletions
|
@ -183,9 +183,9 @@ public final class BinaryDictionary extends Dictionary {
|
|||
private static native void getHeaderInfoNative(long dict, int[] outHeaderSize,
|
||||
int[] outFormatVersion, ArrayList<int[]> outAttributeKeys,
|
||||
ArrayList<int[]> outAttributeValues);
|
||||
private static native void flushNative(long dict, String filePath);
|
||||
private static native boolean flushNative(long dict, String filePath);
|
||||
private static native boolean needsToRunGCNative(long dict, boolean mindsBlockByGC);
|
||||
private static native void flushWithGCNative(long dict, String filePath);
|
||||
private static native boolean flushWithGCNative(long dict, String filePath);
|
||||
private static native void closeNative(long dict);
|
||||
private static native int getFormatVersionNative(long dict);
|
||||
private static native int getProbabilityNative(long dict, int[] word);
|
||||
|
@ -203,12 +203,12 @@ public final class BinaryDictionary extends Dictionary {
|
|||
int[] outputSuggestionCount, int[] outputCodePoints, int[] outputScores,
|
||||
int[] outputIndices, int[] outputTypes, int[] outputAutoCommitFirstWordConfidence,
|
||||
float[] inOutLanguageWeight);
|
||||
private static native void addUnigramWordNative(long dict, int[] word, int probability,
|
||||
private static native boolean addUnigramWordNative(long dict, int[] word, int probability,
|
||||
int[] shortcutTarget, int shortcutProbability, boolean isBeginningOfSentence,
|
||||
boolean isNotAWord, boolean isBlacklisted, int timestamp);
|
||||
private static native void addBigramWordsNative(long dict, int[] word0,
|
||||
private static native boolean addBigramWordsNative(long dict, int[] word0,
|
||||
boolean isBeginningOfSentence, int[] word1, int probability, int timestamp);
|
||||
private static native void removeBigramWordsNative(long dict, int[] word0,
|
||||
private static native boolean removeBigramWordsNative(long dict, int[] word0,
|
||||
boolean isBeginningOfSentence, int[] word1);
|
||||
private static native int addMultipleDictionaryEntriesNative(long dict,
|
||||
LanguageModelParam[] languageModelParams, int startIndex);
|
||||
|
@ -417,45 +417,53 @@ public final class BinaryDictionary extends Dictionary {
|
|||
}
|
||||
|
||||
// Add a unigram entry to binary dictionary with unigram attributes in native code.
|
||||
public void addUnigramEntry(final String word, final int probability,
|
||||
public boolean addUnigramEntry(final String word, final int probability,
|
||||
final String shortcutTarget, final int shortcutProbability,
|
||||
final boolean isBeginningOfSentence, final boolean isNotAWord,
|
||||
final boolean isBlacklisted, final int timestamp) {
|
||||
if (word == null || (word.isEmpty() && !isBeginningOfSentence)) {
|
||||
return;
|
||||
return false;
|
||||
}
|
||||
final int[] codePoints = StringUtils.toCodePointArray(word);
|
||||
final int[] shortcutTargetCodePoints = (shortcutTarget != null) ?
|
||||
StringUtils.toCodePointArray(shortcutTarget) : null;
|
||||
addUnigramWordNative(mNativeDict, codePoints, probability, shortcutTargetCodePoints,
|
||||
shortcutProbability, isBeginningOfSentence, isNotAWord, isBlacklisted, timestamp);
|
||||
if (!addUnigramWordNative(mNativeDict, codePoints, probability, shortcutTargetCodePoints,
|
||||
shortcutProbability, isBeginningOfSentence, isNotAWord, isBlacklisted, timestamp)) {
|
||||
return false;
|
||||
}
|
||||
mHasUpdated = true;
|
||||
return true;
|
||||
}
|
||||
|
||||
// Add an n-gram entry to the binary dictionary with timestamp in native code.
|
||||
public void addNgramEntry(final PrevWordsInfo prevWordsInfo, final String word,
|
||||
final int probability,
|
||||
final int timestamp) {
|
||||
public boolean addNgramEntry(final PrevWordsInfo prevWordsInfo, final String word,
|
||||
final int probability, final int timestamp) {
|
||||
if (!prevWordsInfo.isValid() || TextUtils.isEmpty(word)) {
|
||||
return;
|
||||
return false;
|
||||
}
|
||||
final int[] codePoints0 = StringUtils.toCodePointArray(prevWordsInfo.mPrevWord);
|
||||
final int[] codePoints1 = StringUtils.toCodePointArray(word);
|
||||
addBigramWordsNative(mNativeDict, codePoints0, prevWordsInfo.mIsBeginningOfSentence,
|
||||
codePoints1, probability, timestamp);
|
||||
if (!addBigramWordsNative(mNativeDict, codePoints0, prevWordsInfo.mIsBeginningOfSentence,
|
||||
codePoints1, probability, timestamp)) {
|
||||
return false;
|
||||
}
|
||||
mHasUpdated = true;
|
||||
return true;
|
||||
}
|
||||
|
||||
// Remove an n-gram entry from the binary dictionary in native code.
|
||||
public void removeNgramEntry(final PrevWordsInfo prevWordsInfo, final String word) {
|
||||
public boolean removeNgramEntry(final PrevWordsInfo prevWordsInfo, final String word) {
|
||||
if (!prevWordsInfo.isValid() || TextUtils.isEmpty(word)) {
|
||||
return;
|
||||
return false;
|
||||
}
|
||||
final int[] codePoints0 = StringUtils.toCodePointArray(prevWordsInfo.mPrevWord);
|
||||
final int[] codePoints1 = StringUtils.toCodePointArray(word);
|
||||
removeBigramWordsNative(mNativeDict, codePoints0, prevWordsInfo.mIsBeginningOfSentence,
|
||||
codePoints1);
|
||||
if (!removeBigramWordsNative(mNativeDict, codePoints0, prevWordsInfo.mIsBeginningOfSentence,
|
||||
codePoints1)) {
|
||||
return false;
|
||||
}
|
||||
mHasUpdated = true;
|
||||
return true;
|
||||
}
|
||||
|
||||
public void addMultipleDictionaryEntries(final LanguageModelParam[] languageModelParams) {
|
||||
|
@ -485,26 +493,33 @@ public final class BinaryDictionary extends Dictionary {
|
|||
}
|
||||
|
||||
// Flush to dict file if the dictionary has been updated.
|
||||
public void flush() {
|
||||
if (!isValidDictionary()) return;
|
||||
public boolean flush() {
|
||||
if (!isValidDictionary()) return false;
|
||||
if (mHasUpdated) {
|
||||
flushNative(mNativeDict, mDictFilePath);
|
||||
if (!flushNative(mNativeDict, mDictFilePath)) {
|
||||
return false;
|
||||
}
|
||||
reopen();
|
||||
}
|
||||
return true;
|
||||
}
|
||||
|
||||
// Run GC and flush to dict file if the dictionary has been updated.
|
||||
public void flushWithGCIfHasUpdated() {
|
||||
public boolean flushWithGCIfHasUpdated() {
|
||||
if (mHasUpdated) {
|
||||
flushWithGC();
|
||||
return flushWithGC();
|
||||
}
|
||||
return true;
|
||||
}
|
||||
|
||||
// Run GC and flush to dict file.
|
||||
public void flushWithGC() {
|
||||
if (!isValidDictionary()) return;
|
||||
flushWithGCNative(mNativeDict, mDictFilePath);
|
||||
public boolean flushWithGC() {
|
||||
if (!isValidDictionary()) return false;
|
||||
if (!flushWithGCNative(mNativeDict, mDictFilePath)) {
|
||||
return false;
|
||||
}
|
||||
reopen();
|
||||
return true;
|
||||
}
|
||||
|
||||
/**
|
||||
|
|
|
@ -298,8 +298,10 @@ abstract public class ExpandableBinaryDictionary extends Dictionary {
|
|||
protected void addUnigramLocked(final String word, final int frequency,
|
||||
final String shortcutTarget, final int shortcutFreq, final boolean isNotAWord,
|
||||
final boolean isBlacklisted, final int timestamp) {
|
||||
mBinaryDictionary.addUnigramEntry(word, frequency, shortcutTarget, shortcutFreq,
|
||||
false /* isBeginningOfSentence */, isNotAWord, isBlacklisted, timestamp);
|
||||
if (!mBinaryDictionary.addUnigramEntry(word, frequency, shortcutTarget, shortcutFreq,
|
||||
false /* isBeginningOfSentence */, isNotAWord, isBlacklisted, timestamp)) {
|
||||
Log.e(TAG, "Cannot add unigram entry. word: " + word);
|
||||
}
|
||||
}
|
||||
|
||||
/**
|
||||
|
@ -322,7 +324,11 @@ abstract public class ExpandableBinaryDictionary extends Dictionary {
|
|||
|
||||
protected void addNgramEntryLocked(final PrevWordsInfo prevWordsInfo, final String word,
|
||||
final int frequency, final int timestamp) {
|
||||
mBinaryDictionary.addNgramEntry(prevWordsInfo, word, frequency, timestamp);
|
||||
if (!mBinaryDictionary.addNgramEntry(prevWordsInfo, word, frequency, timestamp)) {
|
||||
Log.e(TAG, "Cannot add n-gram entry.");
|
||||
Log.e(TAG, " PrevWordsInfo: " + prevWordsInfo);
|
||||
Log.e(TAG, " word: " + word);
|
||||
}
|
||||
}
|
||||
|
||||
/**
|
||||
|
|
|
@ -53,4 +53,10 @@ public class PrevWordsInfo {
|
|||
public boolean isValid() {
|
||||
return mPrevWord != null;
|
||||
}
|
||||
|
||||
@Override
|
||||
public String toString() {
|
||||
return "PrevWord: " + mPrevWord + ", isBeginningOfSentence: "
|
||||
+ mIsBeginningOfSentence + ".";
|
||||
}
|
||||
}
|
||||
|
|
|
@ -95,15 +95,15 @@ static jlong latinime_BinaryDictionary_createOnMemory(JNIEnv *env, jclass clazz,
|
|||
return reinterpret_cast<jlong>(dictionary);
|
||||
}
|
||||
|
||||
static void latinime_BinaryDictionary_flush(JNIEnv *env, jclass clazz, jlong dict,
|
||||
static bool latinime_BinaryDictionary_flush(JNIEnv *env, jclass clazz, jlong dict,
|
||||
jstring filePath) {
|
||||
Dictionary *dictionary = reinterpret_cast<Dictionary *>(dict);
|
||||
if (!dictionary) return;
|
||||
if (!dictionary) return false;
|
||||
const jsize filePathUtf8Length = env->GetStringUTFLength(filePath);
|
||||
char filePathChars[filePathUtf8Length + 1];
|
||||
env->GetStringUTFRegion(filePath, 0, env->GetStringLength(filePath), filePathChars);
|
||||
filePathChars[filePathUtf8Length] = '\0';
|
||||
dictionary->flush(filePathChars);
|
||||
return dictionary->flush(filePathChars);
|
||||
}
|
||||
|
||||
static bool latinime_BinaryDictionary_needsToRunGC(JNIEnv *env, jclass clazz,
|
||||
|
@ -113,15 +113,15 @@ static bool latinime_BinaryDictionary_needsToRunGC(JNIEnv *env, jclass clazz,
|
|||
return dictionary->needsToRunGC(mindsBlockByGC == JNI_TRUE);
|
||||
}
|
||||
|
||||
static void latinime_BinaryDictionary_flushWithGC(JNIEnv *env, jclass clazz, jlong dict,
|
||||
static bool latinime_BinaryDictionary_flushWithGC(JNIEnv *env, jclass clazz, jlong dict,
|
||||
jstring filePath) {
|
||||
Dictionary *dictionary = reinterpret_cast<Dictionary *>(dict);
|
||||
if (!dictionary) return;
|
||||
if (!dictionary) return false;
|
||||
const jsize filePathUtf8Length = env->GetStringUTFLength(filePath);
|
||||
char filePathChars[filePathUtf8Length + 1];
|
||||
env->GetStringUTFRegion(filePath, 0, env->GetStringLength(filePath), filePathChars);
|
||||
filePathChars[filePathUtf8Length] = '\0';
|
||||
dictionary->flushWithGC(filePathChars);
|
||||
return dictionary->flushWithGC(filePathChars);
|
||||
}
|
||||
|
||||
static void latinime_BinaryDictionary_close(JNIEnv *env, jclass clazz, jlong dict) {
|
||||
|
@ -324,13 +324,13 @@ static void latinime_BinaryDictionary_getWordProperty(JNIEnv *env, jclass clazz,
|
|||
outShortcutProbabilities);
|
||||
}
|
||||
|
||||
static void latinime_BinaryDictionary_addUnigramWord(JNIEnv *env, jclass clazz, jlong dict,
|
||||
static bool latinime_BinaryDictionary_addUnigramWord(JNIEnv *env, jclass clazz, jlong dict,
|
||||
jintArray word, jint probability, jintArray shortcutTarget, jint shortcutProbability,
|
||||
jboolean isBeginningOfSentence, jboolean isNotAWord, jboolean isBlacklisted,
|
||||
jint timestamp) {
|
||||
Dictionary *dictionary = reinterpret_cast<Dictionary *>(dict);
|
||||
if (!dictionary) {
|
||||
return;
|
||||
return false;
|
||||
}
|
||||
jsize codePointCount = env->GetArrayLength(word);
|
||||
int codePoints[codePointCount];
|
||||
|
@ -344,15 +344,15 @@ static void latinime_BinaryDictionary_addUnigramWord(JNIEnv *env, jclass clazz,
|
|||
// Use 1 for count to indicate the word has inputted.
|
||||
const UnigramProperty unigramProperty(isBeginningOfSentence, isNotAWord,
|
||||
isBlacklisted, probability, timestamp, 0 /* level */, 1 /* count */, &shortcuts);
|
||||
dictionary->addUnigramEntry(codePoints, codePointCount, &unigramProperty);
|
||||
return dictionary->addUnigramEntry(codePoints, codePointCount, &unigramProperty);
|
||||
}
|
||||
|
||||
static void latinime_BinaryDictionary_addBigramWords(JNIEnv *env, jclass clazz, jlong dict,
|
||||
static bool latinime_BinaryDictionary_addBigramWords(JNIEnv *env, jclass clazz, jlong dict,
|
||||
jintArray word0, jboolean isBeginningOfSentence, jintArray word1, jint probability,
|
||||
jint timestamp) {
|
||||
Dictionary *dictionary = reinterpret_cast<Dictionary *>(dict);
|
||||
if (!dictionary) {
|
||||
return;
|
||||
return false;
|
||||
}
|
||||
jsize word0Length = env->GetArrayLength(word0);
|
||||
int word0CodePoints[word0Length];
|
||||
|
@ -366,14 +366,14 @@ static void latinime_BinaryDictionary_addBigramWords(JNIEnv *env, jclass clazz,
|
|||
const BigramProperty bigramProperty(&bigramTargetCodePoints, probability,
|
||||
timestamp, 0 /* level */, 1 /* count */);
|
||||
const PrevWordsInfo prevWordsInfo(word0CodePoints, word0Length, isBeginningOfSentence);
|
||||
dictionary->addNgramEntry(&prevWordsInfo, &bigramProperty);
|
||||
return dictionary->addNgramEntry(&prevWordsInfo, &bigramProperty);
|
||||
}
|
||||
|
||||
static void latinime_BinaryDictionary_removeBigramWords(JNIEnv *env, jclass clazz, jlong dict,
|
||||
static bool latinime_BinaryDictionary_removeBigramWords(JNIEnv *env, jclass clazz, jlong dict,
|
||||
jintArray word0, jboolean isBeginningOfSentence, jintArray word1) {
|
||||
Dictionary *dictionary = reinterpret_cast<Dictionary *>(dict);
|
||||
if (!dictionary) {
|
||||
return;
|
||||
return false;
|
||||
}
|
||||
jsize word0Length = env->GetArrayLength(word0);
|
||||
int word0CodePoints[word0Length];
|
||||
|
@ -382,7 +382,7 @@ static void latinime_BinaryDictionary_removeBigramWords(JNIEnv *env, jclass claz
|
|||
int word1CodePoints[word1Length];
|
||||
env->GetIntArrayRegion(word1, 0, word1Length, word1CodePoints);
|
||||
const PrevWordsInfo prevWordsInfo(word0CodePoints, word0Length, isBeginningOfSentence);
|
||||
dictionary->removeNgramEntry(&prevWordsInfo, word1CodePoints, word1Length);
|
||||
return dictionary->removeNgramEntry(&prevWordsInfo, word1CodePoints, word1Length);
|
||||
}
|
||||
|
||||
// Returns how many language model params are processed.
|
||||
|
@ -610,7 +610,7 @@ static const JNINativeMethod sMethods[] = {
|
|||
},
|
||||
{
|
||||
const_cast<char *>("flushNative"),
|
||||
const_cast<char *>("(JLjava/lang/String;)V"),
|
||||
const_cast<char *>("(JLjava/lang/String;)Z"),
|
||||
reinterpret_cast<void *>(latinime_BinaryDictionary_flush)
|
||||
},
|
||||
{
|
||||
|
@ -620,7 +620,7 @@ static const JNINativeMethod sMethods[] = {
|
|||
},
|
||||
{
|
||||
const_cast<char *>("flushWithGCNative"),
|
||||
const_cast<char *>("(JLjava/lang/String;)V"),
|
||||
const_cast<char *>("(JLjava/lang/String;)Z"),
|
||||
reinterpret_cast<void *>(latinime_BinaryDictionary_flushWithGC)
|
||||
},
|
||||
{
|
||||
|
@ -651,17 +651,17 @@ static const JNINativeMethod sMethods[] = {
|
|||
},
|
||||
{
|
||||
const_cast<char *>("addUnigramWordNative"),
|
||||
const_cast<char *>("(J[II[IIZZZI)V"),
|
||||
const_cast<char *>("(J[II[IIZZZI)Z"),
|
||||
reinterpret_cast<void *>(latinime_BinaryDictionary_addUnigramWord)
|
||||
},
|
||||
{
|
||||
const_cast<char *>("addBigramWordsNative"),
|
||||
const_cast<char *>("(J[IZ[III)V"),
|
||||
const_cast<char *>("(J[IZ[III)Z"),
|
||||
reinterpret_cast<void *>(latinime_BinaryDictionary_addBigramWords)
|
||||
},
|
||||
{
|
||||
const_cast<char *>("removeBigramWordsNative"),
|
||||
const_cast<char *>("(J[IZ[I)V"),
|
||||
const_cast<char *>("(J[IZ[I)Z"),
|
||||
reinterpret_cast<void *>(latinime_BinaryDictionary_removeBigramWords)
|
||||
},
|
||||
{
|
||||
|
|
|
@ -80,38 +80,38 @@ int Dictionary::getBigramProbability(const PrevWordsInfo *const prevWordsInfo, c
|
|||
return mBigramDictionary.getBigramProbability(prevWordsInfo, word, length);
|
||||
}
|
||||
|
||||
void Dictionary::addUnigramEntry(const int *const word, const int length,
|
||||
bool Dictionary::addUnigramEntry(const int *const word, const int length,
|
||||
const UnigramProperty *const unigramProperty) {
|
||||
if (unigramProperty->representsBeginningOfSentence()
|
||||
&& !mDictionaryStructureWithBufferPolicy->getHeaderStructurePolicy()
|
||||
->supportsBeginningOfSentence()) {
|
||||
AKLOGE("The dictionary doesn't support Beginning-of-Sentence.");
|
||||
return;
|
||||
return false;
|
||||
}
|
||||
TimeKeeper::setCurrentTime();
|
||||
mDictionaryStructureWithBufferPolicy->addUnigramEntry(word, length, unigramProperty);
|
||||
return mDictionaryStructureWithBufferPolicy->addUnigramEntry(word, length, unigramProperty);
|
||||
}
|
||||
|
||||
void Dictionary::addNgramEntry(const PrevWordsInfo *const prevWordsInfo,
|
||||
bool Dictionary::addNgramEntry(const PrevWordsInfo *const prevWordsInfo,
|
||||
const BigramProperty *const bigramProperty) {
|
||||
TimeKeeper::setCurrentTime();
|
||||
mDictionaryStructureWithBufferPolicy->addNgramEntry(prevWordsInfo, bigramProperty);
|
||||
return mDictionaryStructureWithBufferPolicy->addNgramEntry(prevWordsInfo, bigramProperty);
|
||||
}
|
||||
|
||||
void Dictionary::removeNgramEntry(const PrevWordsInfo *const prevWordsInfo,
|
||||
bool Dictionary::removeNgramEntry(const PrevWordsInfo *const prevWordsInfo,
|
||||
const int *const word, const int length) {
|
||||
TimeKeeper::setCurrentTime();
|
||||
mDictionaryStructureWithBufferPolicy->removeNgramEntry(prevWordsInfo, word, length);
|
||||
return mDictionaryStructureWithBufferPolicy->removeNgramEntry(prevWordsInfo, word, length);
|
||||
}
|
||||
|
||||
void Dictionary::flush(const char *const filePath) {
|
||||
bool Dictionary::flush(const char *const filePath) {
|
||||
TimeKeeper::setCurrentTime();
|
||||
mDictionaryStructureWithBufferPolicy->flush(filePath);
|
||||
return mDictionaryStructureWithBufferPolicy->flush(filePath);
|
||||
}
|
||||
|
||||
void Dictionary::flushWithGC(const char *const filePath) {
|
||||
bool Dictionary::flushWithGC(const char *const filePath) {
|
||||
TimeKeeper::setCurrentTime();
|
||||
mDictionaryStructureWithBufferPolicy->flushWithGC(filePath);
|
||||
return mDictionaryStructureWithBufferPolicy->flushWithGC(filePath);
|
||||
}
|
||||
|
||||
bool Dictionary::needsToRunGC(const bool mindsBlockByGC) {
|
||||
|
|
|
@ -76,18 +76,18 @@ class Dictionary {
|
|||
int getBigramProbability(const PrevWordsInfo *const prevWordsInfo,
|
||||
const int *word, int length) const;
|
||||
|
||||
void addUnigramEntry(const int *const codePoints, const int codePointCount,
|
||||
bool addUnigramEntry(const int *const codePoints, const int codePointCount,
|
||||
const UnigramProperty *const unigramProperty);
|
||||
|
||||
void addNgramEntry(const PrevWordsInfo *const prevWordsInfo,
|
||||
bool addNgramEntry(const PrevWordsInfo *const prevWordsInfo,
|
||||
const BigramProperty *const bigramProperty);
|
||||
|
||||
void removeNgramEntry(const PrevWordsInfo *const prevWordsInfo, const int *const word,
|
||||
bool removeNgramEntry(const PrevWordsInfo *const prevWordsInfo, const int *const word,
|
||||
const int length);
|
||||
|
||||
void flush(const char *const filePath);
|
||||
bool flush(const char *const filePath);
|
||||
|
||||
void flushWithGC(const char *const filePath);
|
||||
bool flushWithGC(const char *const filePath);
|
||||
|
||||
bool needsToRunGC(const bool mindsBlockByGC);
|
||||
|
||||
|
|
|
@ -81,9 +81,11 @@ class DictionaryStructureWithBufferPolicy {
|
|||
virtual bool removeNgramEntry(const PrevWordsInfo *const prevWordsInfo,
|
||||
const int *const word, const int length) = 0;
|
||||
|
||||
virtual void flush(const char *const filePath) = 0;
|
||||
// Returns whether the flush was success or not.
|
||||
virtual bool flush(const char *const filePath) = 0;
|
||||
|
||||
virtual void flushWithGC(const char *const filePath) = 0;
|
||||
// Returns whether the GC and flush were success or not.
|
||||
virtual bool flushWithGC(const char *const filePath) = 0;
|
||||
|
||||
virtual bool needsToRunGC(const bool mindsBlockByGC) const = 0;
|
||||
|
||||
|
|
|
@ -296,26 +296,30 @@ bool Ver4PatriciaTriePolicy::removeNgramEntry(const PrevWordsInfo *const prevWor
|
|||
}
|
||||
}
|
||||
|
||||
void Ver4PatriciaTriePolicy::flush(const char *const filePath) {
|
||||
bool Ver4PatriciaTriePolicy::flush(const char *const filePath) {
|
||||
if (!mBuffers->isUpdatable()) {
|
||||
AKLOGI("Warning: flush() is called for non-updatable dictionary. filePath: %s", filePath);
|
||||
return;
|
||||
return false;
|
||||
}
|
||||
if (!mWritingHelper.writeToDictFile(filePath, mUnigramCount, mBigramCount)) {
|
||||
AKLOGE("Cannot flush the dictionary to file.");
|
||||
mIsCorrupted = true;
|
||||
return false;
|
||||
}
|
||||
return true;
|
||||
}
|
||||
|
||||
void Ver4PatriciaTriePolicy::flushWithGC(const char *const filePath) {
|
||||
bool Ver4PatriciaTriePolicy::flushWithGC(const char *const filePath) {
|
||||
if (!mBuffers->isUpdatable()) {
|
||||
AKLOGI("Warning: flushWithGC() is called for non-updatable dictionary.");
|
||||
return;
|
||||
return false;
|
||||
}
|
||||
if (!mWritingHelper.writeToDictFileWithGC(getRootPosition(), filePath)) {
|
||||
AKLOGE("Cannot flush the dictionary to file with GC.");
|
||||
mIsCorrupted = true;
|
||||
return false;
|
||||
}
|
||||
return true;
|
||||
}
|
||||
|
||||
bool Ver4PatriciaTriePolicy::needsToRunGC(const bool mindsBlockByGC) const {
|
||||
|
|
|
@ -117,9 +117,9 @@ class Ver4PatriciaTriePolicy : public DictionaryStructureWithBufferPolicy {
|
|||
bool removeNgramEntry(const PrevWordsInfo *const prevWordsInfo, const int *const word,
|
||||
const int length);
|
||||
|
||||
void flush(const char *const filePath);
|
||||
bool flush(const char *const filePath);
|
||||
|
||||
void flushWithGC(const char *const filePath);
|
||||
bool flushWithGC(const char *const filePath);
|
||||
|
||||
bool needsToRunGC(const bool mindsBlockByGC) const;
|
||||
|
||||
|
|
|
@ -102,14 +102,16 @@ class PatriciaTriePolicy : public DictionaryStructureWithBufferPolicy {
|
|||
return false;
|
||||
}
|
||||
|
||||
void flush(const char *const filePath) {
|
||||
bool flush(const char *const filePath) {
|
||||
// This method should not be called for non-updatable dictionary.
|
||||
AKLOGI("Warning: flush() is called for non-updatable dictionary.");
|
||||
return false;
|
||||
}
|
||||
|
||||
void flushWithGC(const char *const filePath) {
|
||||
bool flushWithGC(const char *const filePath) {
|
||||
// This method should not be called for non-updatable dictionary.
|
||||
AKLOGI("Warning: flushWithGC() is called for non-updatable dictionary.");
|
||||
return false;
|
||||
}
|
||||
|
||||
bool needsToRunGC(const bool mindsBlockByGC) const {
|
||||
|
|
|
@ -304,26 +304,30 @@ bool Ver4PatriciaTriePolicy::removeNgramEntry(const PrevWordsInfo *const prevWor
|
|||
}
|
||||
}
|
||||
|
||||
void Ver4PatriciaTriePolicy::flush(const char *const filePath) {
|
||||
bool Ver4PatriciaTriePolicy::flush(const char *const filePath) {
|
||||
if (!mBuffers->isUpdatable()) {
|
||||
AKLOGI("Warning: flush() is called for non-updatable dictionary. filePath: %s", filePath);
|
||||
return;
|
||||
return false;
|
||||
}
|
||||
if (!mWritingHelper.writeToDictFile(filePath, mUnigramCount, mBigramCount)) {
|
||||
AKLOGE("Cannot flush the dictionary to file.");
|
||||
mIsCorrupted = true;
|
||||
return false;
|
||||
}
|
||||
return true;
|
||||
}
|
||||
|
||||
void Ver4PatriciaTriePolicy::flushWithGC(const char *const filePath) {
|
||||
bool Ver4PatriciaTriePolicy::flushWithGC(const char *const filePath) {
|
||||
if (!mBuffers->isUpdatable()) {
|
||||
AKLOGI("Warning: flushWithGC() is called for non-updatable dictionary.");
|
||||
return;
|
||||
return false;
|
||||
}
|
||||
if (!mWritingHelper.writeToDictFileWithGC(getRootPosition(), filePath)) {
|
||||
AKLOGE("Cannot flush the dictionary to file with GC.");
|
||||
mIsCorrupted = true;
|
||||
return false;
|
||||
}
|
||||
return true;
|
||||
}
|
||||
|
||||
bool Ver4PatriciaTriePolicy::needsToRunGC(const bool mindsBlockByGC) const {
|
||||
|
|
|
@ -99,9 +99,9 @@ class Ver4PatriciaTriePolicy : public DictionaryStructureWithBufferPolicy {
|
|||
bool removeNgramEntry(const PrevWordsInfo *const prevWordsInfo, const int *const word1,
|
||||
const int length1);
|
||||
|
||||
void flush(const char *const filePath);
|
||||
bool flush(const char *const filePath);
|
||||
|
||||
void flushWithGC(const char *const filePath);
|
||||
bool flushWithGC(const char *const filePath);
|
||||
|
||||
bool needsToRunGC(const bool mindsBlockByGC) const;
|
||||
|
||||
|
|
|
@ -75,33 +75,54 @@ public class Ver4DictEncoder implements DictEncoder {
|
|||
for (final WordProperty wordProperty : dict) {
|
||||
// TODO: switch to addMultipleDictionaryEntries when they support shortcuts
|
||||
if (null == wordProperty.mShortcutTargets || wordProperty.mShortcutTargets.isEmpty()) {
|
||||
binaryDict.addUnigramEntry(wordProperty.mWord, wordProperty.getProbability(),
|
||||
if (!binaryDict.addUnigramEntry(wordProperty.mWord, wordProperty.getProbability(),
|
||||
null /* shortcutTarget */, 0 /* shortcutProbability */,
|
||||
wordProperty.mIsBeginningOfSentence, wordProperty.mIsNotAWord,
|
||||
wordProperty.mIsBlacklistEntry, 0 /* timestamp */);
|
||||
wordProperty.mIsBlacklistEntry, 0 /* timestamp */)) {
|
||||
MakedictLog.e("Cannot add unigram entry for " + wordProperty.mWord);
|
||||
}
|
||||
} else {
|
||||
for (final WeightedString shortcutTarget : wordProperty.mShortcutTargets) {
|
||||
binaryDict.addUnigramEntry(wordProperty.mWord, wordProperty.getProbability(),
|
||||
if (!binaryDict.addUnigramEntry(wordProperty.mWord,
|
||||
wordProperty.getProbability(),
|
||||
shortcutTarget.mWord, shortcutTarget.getProbability(),
|
||||
wordProperty.mIsBeginningOfSentence, wordProperty.mIsNotAWord,
|
||||
wordProperty.mIsBlacklistEntry, 0 /* timestamp */);
|
||||
wordProperty.mIsBlacklistEntry, 0 /* timestamp */)) {
|
||||
MakedictLog.e("Cannot add unigram entry for " + wordProperty.mWord
|
||||
+ ", shortcutTarget: " + shortcutTarget.mWord);
|
||||
return;
|
||||
}
|
||||
}
|
||||
}
|
||||
if (binaryDict.needsToRunGC(true /* mindsBlockByGC */)) {
|
||||
binaryDict.flushWithGC();
|
||||
if (!binaryDict.flushWithGC()) {
|
||||
MakedictLog.e("Cannot flush dict with GC.");
|
||||
return;
|
||||
}
|
||||
}
|
||||
}
|
||||
for (final WordProperty word0Property : dict) {
|
||||
if (null == word0Property.mBigrams) continue;
|
||||
for (final WeightedString word1 : word0Property.mBigrams) {
|
||||
binaryDict.addNgramEntry(new PrevWordsInfo(word0Property.mWord), word1.mWord,
|
||||
word1.getProbability(), 0 /* timestamp */);
|
||||
final PrevWordsInfo prevWordsInfo = new PrevWordsInfo(word0Property.mWord);
|
||||
if (!binaryDict.addNgramEntry(prevWordsInfo, word1.mWord,
|
||||
word1.getProbability(), 0 /* timestamp */)) {
|
||||
MakedictLog.e("Cannot add n-gram entry for "
|
||||
+ prevWordsInfo + " -> " + word1.mWord);
|
||||
return;
|
||||
}
|
||||
if (binaryDict.needsToRunGC(true /* mindsBlockByGC */)) {
|
||||
binaryDict.flushWithGC();
|
||||
if (!binaryDict.flushWithGC()) {
|
||||
MakedictLog.e("Cannot flush dict with GC.");
|
||||
return;
|
||||
}
|
||||
}
|
||||
}
|
||||
binaryDict.flushWithGC();
|
||||
}
|
||||
if (!binaryDict.flushWithGC()) {
|
||||
MakedictLog.e("Cannot flush dict with GC.");
|
||||
return;
|
||||
}
|
||||
binaryDict.close();
|
||||
}
|
||||
|
||||
|
|
Loading…
Reference in a new issue