Skip to content
Snippets Groups Projects
Commit d4952c8f authored by satok's avatar satok
Browse files

Move a logic for finding words with a missing character to the native code. 

Change-Id: I58338643830ff4f9708f78a9c26f75c8bf2ebf45
parent bd20db25
No related branches found
No related tags found
No related merge requests found
Hide whitespace changes
Inline Side-by-side
java/src/com/android/inputmethod/latin/BinaryDictionary.java
+ 2
− 19
View file @ d4952c8f
......@@ -45,7 +45,6 @@ public class BinaryDictionary extends Dictionary {
private static final int MAX_BIGRAMS = 60;
private static final int TYPED_LETTER_MULTIPLIER = 2;
private static final boolean ENABLE_MISSED_CHARACTERS = true;
private int mDicTypeId;
private int mNativeDict;
......@@ -199,27 +198,11 @@ public class BinaryDictionary extends Dictionary {
Arrays.fill(mOutputChars, (char) 0);
Arrays.fill(mFrequencies, 0);
int count = getSuggestionsNative(mNativeDict, mInputCodes, codesSize,
mOutputChars, mFrequencies,
MAX_WORD_LENGTH, MAX_WORDS, MAX_ALTERNATIVES, -1,
int count = getSuggestionsNative(mNativeDict, mInputCodes, codesSize, mOutputChars,
mFrequencies, MAX_WORD_LENGTH, MAX_WORDS, MAX_ALTERNATIVES, -1,
nextLettersFrequencies,
nextLettersFrequencies != null ? nextLettersFrequencies.length : 0);
// If there aren't sufficient suggestions, search for words by allowing wild cards at
// the different character positions. This feature is not ready for prime-time as we need
// to figure out the best ranking for such words compared to proximity corrections and
// completions.
if (ENABLE_MISSED_CHARACTERS && count < 5) {
for (int skip = 0; skip < codesSize; skip++) {
int tempCount = getSuggestionsNative(mNativeDict, mInputCodes, codesSize,
mOutputChars, mFrequencies,
MAX_WORD_LENGTH, MAX_WORDS, MAX_ALTERNATIVES, skip,
null, 0);
count = Math.max(count, tempCount);
if (tempCount > 0) break;
}
}
for (int j = 0; j < count; j++) {
if (mFrequencies[j] < 1) break;
int start = j * MAX_WORD_LENGTH;
......
native/src/dictionary.cpp
+ 66
− 35
View file @ d4952c8f
......@@ -37,6 +37,10 @@
#define DICTIONARY_HEADER_SIZE 2
#define NOT_VALID_WORD -99
#define SUGGEST_MISSING_CHARACTERS true
#define SUGGEST_MISSING_CHARACTERS_THRESHOLD 5
namespace latinime {
Dictionary::Dictionary(void *dict, int typedLetterMultiplier, int fullWordMultiplier)
......@@ -56,7 +60,42 @@ int Dictionary::getSuggestions(int *codes, int codesSize, unsigned short *outWor
int maxWordLength, int maxWords, int maxAlternatives, int skipPos,
int *nextLetters, int nextLettersSize)
{
int suggWords;
initSuggestions(codes, codesSize, outWords, frequencies, maxWordLength, maxWords,
maxAlternatives);
int suggestedWordsCount = getSuggestionCandidates(codesSize, maxWords, skipPos, nextLetters,
nextLettersSize);
// If there aren't sufficient suggestions, search for words by allowing wild cards at
// the different character positions. This feature is not ready for prime-time as we need
// to figure out the best ranking for such words compared to proximity corrections and
// completions.
if (SUGGEST_MISSING_CHARACTERS && suggestedWordsCount < SUGGEST_MISSING_CHARACTERS_THRESHOLD) {
for (int i = 0; i < codesSize; ++i) {
int tempCount = getSuggestionCandidates(codesSize, maxWords, i, NULL, 0);
if (tempCount > suggestedWordsCount) {
suggestedWordsCount = tempCount;
break;
}
}
}
if (DEBUG_DICT) {
LOGI("Returning %d words", suggestedWordsCount);
LOGI("Next letters: ");
for (int k = 0; k < nextLettersSize; k++) {
if (nextLetters[k] > 0) {
LOGI("%c = %d,", k, nextLetters[k]);
}
}
LOGI("\n");
}
return suggestedWordsCount;
}
void Dictionary::initSuggestions(int *codes, int codesSize, unsigned short *outWords,
int *frequencies, int maxWordLength, int maxWords, int maxAlternatives) {
mFrequencies = frequencies;
mOutputChars = outWords;
mInputCodes = codes;
......@@ -64,39 +103,29 @@ int Dictionary::getSuggestions(int *codes, int codesSize, unsigned short *outWor
mMaxAlternatives = maxAlternatives;
mMaxWordLength = maxWordLength;
mMaxWords = maxWords;
mSkipPos = skipPos;
mMaxEditDistance = mInputLength < 5 ? 2 : mInputLength / 2;
mNextLettersFrequencies = nextLetters;
mNextLettersSize = nextLettersSize;
}
int Dictionary::getSuggestionCandidates(int inputLength, int maxWords, int skipPos,
int *nextLetters, int nextLettersSize) {
if (checkIfDictVersionIsLatest()) {
getWordsRec(DICTIONARY_HEADER_SIZE, 0, mInputLength * 3, false, 1, 0, 0);
getWordsRec(DICTIONARY_HEADER_SIZE, 0, inputLength * 3, false, 1, 0, 0, skipPos,
nextLetters, nextLettersSize);
} else {
getWordsRec(0, 0, mInputLength * 3, false, 1, 0, 0);
getWordsRec(0, 0, inputLength * 3, false, 1, 0, 0, skipPos, nextLetters, nextLettersSize);
}
// Get the word count
suggWords = 0;
while (suggWords < mMaxWords && mFrequencies[suggWords] > 0) suggWords++;
if (DEBUG_DICT) LOGI("Returning %d words", suggWords);
if (DEBUG_DICT) {
LOGI("Next letters: ");
for (int k = 0; k < nextLettersSize; k++) {
if (mNextLettersFrequencies[k] > 0) {
LOGI("%c = %d,", k, mNextLettersFrequencies[k]);
}
}
LOGI("\n");
int suggestedWordsCount = 0;
while (suggestedWordsCount < maxWords && mFrequencies[suggestedWordsCount] > 0) {
suggestedWordsCount++;
}
return suggWords;
return suggestedWordsCount;
}
void
Dictionary::registerNextLetter(unsigned short c)
{
if (c < mNextLettersSize) {
mNextLettersFrequencies[c]++;
void Dictionary::registerNextLetter(unsigned short c, int *nextLetters, int nextLettersSize) {
if (c < nextLettersSize) {
nextLetters[c]++;
}
}
......@@ -287,7 +316,7 @@ static char QUOTE = '\'';
void
Dictionary::getWordsRec(int pos, int depth, int maxDepth, bool completion, int snr, int inputIndex,
int diffs)
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) {
......@@ -321,19 +350,20 @@ Dictionary::getWordsRec(int pos, int depth, int maxDepth, bool completion, int s
mWord[depth] = c;
if (terminal) {
addWord(mWord, depth + 1, freq * snr);
if (depth >= mInputLength && mSkipPos < 0) {
registerNextLetter(mWord[mInputLength]);
if (depth >= mInputLength && skipPos < 0) {
registerNextLetter(mWord[mInputLength], nextLetters, nextLettersSize);
}
}
if (childrenAddress != 0) {
getWordsRec(childrenAddress, depth + 1, maxDepth,
completion, snr, inputIndex, diffs);
getWordsRec(childrenAddress, depth + 1, maxDepth, completion, snr, inputIndex,
diffs, skipPos, nextLetters, nextLettersSize);
}
} else if ((c == QUOTE && currentChars[0] != QUOTE) || mSkipPos == depth) {
} else if ((c == QUOTE && currentChars[0] != QUOTE) || skipPos == depth) {
// Skip the ' or other letter and continue deeper
mWord[depth] = c;
if (childrenAddress != 0) {
getWordsRec(childrenAddress, depth + 1, maxDepth, false, snr, inputIndex, diffs);
getWordsRec(childrenAddress, depth + 1, maxDepth, false, snr, inputIndex, diffs,
skipPos, nextLetters, nextLettersSize);
}
} else {
int j = 0;
......@@ -346,22 +376,23 @@ Dictionary::getWordsRec(int pos, int depth, int maxDepth, bool completion, int s
if (//INCLUDE_TYPED_WORD_IF_VALID ||
!sameAsTyped(mWord, depth + 1)) {
int finalFreq = freq * snr * addedWeight;
if (mSkipPos < 0) finalFreq *= mFullWordMultiplier;
if (skipPos < 0) finalFreq *= mFullWordMultiplier;
addWord(mWord, depth + 1, finalFreq);
}
}
if (childrenAddress != 0) {
getWordsRec(childrenAddress, depth + 1,
maxDepth, true, snr * addedWeight, inputIndex + 1,
diffs + (j > 0));
diffs + (j > 0), skipPos, nextLetters, nextLettersSize);
}
} else if (childrenAddress != 0) {
getWordsRec(childrenAddress, depth + 1, maxDepth,
false, snr * addedWeight, inputIndex + 1, diffs + (j > 0));
false, snr * addedWeight, inputIndex + 1, diffs + (j > 0),
skipPos, nextLetters, nextLettersSize);
}
}
j++;
if (mSkipPos >= 0) break;
if (skipPos >= 0) break;
}
}
}
......
native/src/dictionary.h
+ 6
− 6
View file @ d4952c8f
......@@ -48,7 +48,10 @@ public:
~Dictionary();
private:
void initSuggestions(int *codes, int codesSize, unsigned short *outWords, int *frequencies,
int maxWordLength, int maxWords, int maxAlternatives);
int getSuggestionCandidates(int inputLength, int maxWords, int skipPos, int *nextLetters,
int nextLettersSize);
void getVersionNumber();
bool checkIfDictVersionIsLatest();
int getAddress(int *pos);
......@@ -70,9 +73,9 @@ private:
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 inputIndex, int diffs, int skipPos, int *nextLetters, int nextLettersSize);
int isValidWordRec(int pos, unsigned short *word, int offset, int length);
void registerNextLetter(unsigned short c);
void registerNextLetter(unsigned short c, int *nextLetters, int nextLettersSize);
unsigned char *mDict;
void *mAsset;
......@@ -88,13 +91,10 @@ private:
int mInputLength;
int mMaxAlternatives;
unsigned short mWord[128];
int mSkipPos;
int mMaxEditDistance;
int mFullWordMultiplier;
int mTypedLetterMultiplier;
int *mNextLettersFrequencies;
int mNextLettersSize;
int mVersion;
int mBigram;
};
......
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment