Make WeightedString have ProbabilityInfo. (df1d3e73) · Commits · e / os / android_packages_inputmethods_LatinIME

java/src/com/android/inputmethod/latin/makedict/BinaryDictDecoderUtils.java

+2 −2

Original line number	Diff line number	Diff line
		@@ -511,7 +511,7 @@ public final class BinaryDictDecoderUtils {
		final WeightedString word = getWordAtPosition(dictDecoder, headerSize,
		bigram.mAddress, options);
		final int reconstructedFrequency =
		BinaryDictIOUtils.reconstructBigramFrequency(word.mFrequency,
		BinaryDictIOUtils.reconstructBigramFrequency(word.getProbability(),
		bigram.mFrequency);
		bigrams.add(new WeightedString(word.mWord, reconstructedFrequency));
		}
		@@ -618,7 +618,7 @@ public final class BinaryDictDecoderUtils {
		// words that are not also registered as unigrams so we don't have to avoid
		// them explicitly here.
		for (final WeightedString bigram : w.mBigrams) {
		newDict.setBigram(w.mWord, bigram.mWord, bigram.mFrequency);
		newDict.setBigram(w.mWord, bigram.mWord, bigram.getProbability());
		}
		}
		}

java/src/com/android/inputmethod/latin/makedict/FusionDictionary.java

+26 −15

Original line number	Diff line number	Diff line
		@@ -67,29 +67,40 @@ public final class FusionDictionary implements Iterable<Word> {
		}

		/**
		* A string with a frequency.
		* A string with a probability.
		*
		* This represents an "attribute", that is either a bigram or a shortcut.
		*/
		public static final class WeightedString {
		public final String mWord;
		public int mFrequency;
		public WeightedString(String word, int frequency) {
		public ProbabilityInfo mProbabilityInfo;

		public WeightedString(final String word, final int probability) {
		mWord = word;
		mFrequency = frequency;
		mProbabilityInfo = new ProbabilityInfo(probability);
		}

		public int getProbability() {
		return mProbabilityInfo.mProbability;
		}

		public void setProbability(final int probability) {
		mProbabilityInfo = new ProbabilityInfo(probability);
		}

		@Override
		public int hashCode() {
		return Arrays.hashCode(new Object[] { mWord, mFrequency });
		return Arrays.hashCode(new Object[] { mWord, mProbabilityInfo.mProbability,
		mProbabilityInfo.mTimestamp, mProbabilityInfo.mLevel,
		mProbabilityInfo.mCount });
		}

		@Override
		public boolean equals(Object o) {
		if (o == this) return true;
		if (!(o instanceof WeightedString)) return false;
		WeightedString w = (WeightedString)o;
		return mWord.equals(w.mWord) && mFrequency == w.mFrequency;
		final WeightedString w = (WeightedString)o;
		return mWord.equals(w.mWord) && mProbabilityInfo.equals(w.mProbabilityInfo);
		}
		}

		@@ -200,18 +211,18 @@ public final class FusionDictionary implements Iterable<Word> {
		}

		/**
		* Adds a word to the bigram list. Updates the frequency if the word already
		* Adds a word to the bigram list. Updates the probability if the word already
		* exists.
		*/
		public void addBigram(final String word, final int frequency) {
		public void addBigram(final String word, final int probability) {
		if (mBigrams == null) {
		mBigrams = new ArrayList<WeightedString>();
		}
		WeightedString bigram = getBigram(word);
		if (bigram != null) {
		bigram.mFrequency = frequency;
		bigram.setProbability(probability);
		} else {
		bigram = new WeightedString(word, frequency);
		bigram = new WeightedString(word, probability);
		mBigrams.add(bigram);
		}
		}
		@@ -273,8 +284,8 @@ public final class FusionDictionary implements Iterable<Word> {
		final WeightedString existingShortcut = getShortcut(shortcut.mWord);
		if (existingShortcut == null) {
		mShortcutTargets.add(shortcut);
		} else if (existingShortcut.mFrequency < shortcut.mFrequency) {
		existingShortcut.mFrequency = shortcut.mFrequency;
		} else if (existingShortcut.getProbability() < shortcut.getProbability()) {
		existingShortcut.setProbability(shortcut.getProbability());
		}
		}
		}
		@@ -289,8 +300,8 @@ public final class FusionDictionary implements Iterable<Word> {
		final WeightedString existingBigram = getBigram(bigram.mWord);
		if (existingBigram == null) {
		mBigrams.add(bigram);
		} else if (existingBigram.mFrequency < bigram.mFrequency) {
		existingBigram.mFrequency = bigram.mFrequency;
		} else if (existingBigram.getProbability() < bigram.getProbability()) {
		existingBigram.setProbability(bigram.getProbability());
		}
		}
		}

java/src/com/android/inputmethod/latin/makedict/ProbabilityInfo.java

+17 −0

Original line number	Diff line number	Diff line
		@@ -17,6 +17,7 @@
		package com.android.inputmethod.latin.makedict;

		import com.android.inputmethod.latin.BinaryDictionary;
		import com.android.inputmethod.latin.makedict.FusionDictionary.WeightedString;

		public final class ProbabilityInfo {
		public final int mProbability;
		@@ -39,8 +40,24 @@ public final class ProbabilityInfo {
		mCount = count;
		}

		public boolean hasHistoricalInfo() {
		return mTimestamp != BinaryDictionary.NOT_A_VALID_TIMESTAMP;
		}

		@Override
		public String toString() {
		return mTimestamp + ":" + mLevel + ":" + mCount;
		}

		@Override
		public boolean equals(Object o) {
		if (o == this) return true;
		if (!(o instanceof ProbabilityInfo)) return false;
		final ProbabilityInfo p = (ProbabilityInfo)o;
		if (!hasHistoricalInfo() && !p.hasHistoricalInfo()) {
		return mProbability == p.mProbability;
		}
		return mProbability == p.mProbability && mTimestamp == p.mTimestamp && mLevel == p.mLevel
		&& mCount == p.mCount;
		}
		}
		No newline at end of file

java/src/com/android/inputmethod/latin/makedict/Ver2DictEncoder.java

+2 −2

Original line number	Diff line number	Diff line
		@@ -197,7 +197,7 @@ public class Ver2DictEncoder implements DictEncoder {
		final WeightedString target = shortcutIterator.next();
		final int shortcutFlags = BinaryDictEncoderUtils.makeShortcutFlags(
		shortcutIterator.hasNext(),
		target.mFrequency);
		target.getProbability());
		mPosition = BinaryDictEncoderUtils.writeUIntToBuffer(mBuffer, mPosition, shortcutFlags,
		FormatSpec.PTNODE_ATTRIBUTE_FLAGS_SIZE);
		final int shortcutShift = CharEncoding.writeString(mBuffer, mPosition, target.mWord);
		@@ -231,7 +231,7 @@ public class Ver2DictEncoder implements DictEncoder {
		final int offset = addressOfBigram
		- (mPosition + FormatSpec.PTNODE_ATTRIBUTE_FLAGS_SIZE);
		final int bigramFlags = BinaryDictEncoderUtils.makeBigramFlags(bigramIterator.hasNext(),
		offset, bigram.mFrequency, unigramFrequencyForThisWord, bigram.mWord);
		offset, bigram.getProbability(), unigramFrequencyForThisWord, bigram.mWord);
		mPosition = BinaryDictEncoderUtils.writeUIntToBuffer(mBuffer, mPosition, bigramFlags,
		FormatSpec.PTNODE_ATTRIBUTE_FLAGS_SIZE);
		mPosition += BinaryDictEncoderUtils.writeChildrenPosition(mBuffer, mPosition,

java/src/com/android/inputmethod/latin/makedict/Ver4DictEncoder.java

+2 −2

Original line number	Diff line number	Diff line
		@@ -78,7 +78,7 @@ public class Ver4DictEncoder implements DictEncoder {
		} else {
		for (final WeightedString shortcutTarget : word.mShortcutTargets) {
		binaryDict.addUnigramWord(word.mWord, word.mFrequency,
		shortcutTarget.mWord, shortcutTarget.mFrequency,
		shortcutTarget.mWord, shortcutTarget.getProbability(),
		word.mIsNotAWord, word.mIsBlacklistEntry, 0 /* timestamp */);
		}
		}
		@@ -89,7 +89,7 @@ public class Ver4DictEncoder implements DictEncoder {
		for (final Word word0 : dict) {
		if (null == word0.mBigrams) continue;
		for (final WeightedString word1 : word0.mBigrams) {
		binaryDict.addBigramWords(word0.mWord, word1.mWord, word1.mFrequency,
		binaryDict.addBigramWords(word0.mWord, word1.mWord, word1.getProbability(),
		0 /* timestamp */);
		if (binaryDict.needsToRunGC(true /* mindsBlockByGC */)) {
		binaryDict.flushWithGC();