aboutsummaryrefslogtreecommitdiffstats
path: root/java/src
diff options
context:
space:
mode:
Diffstat (limited to 'java/src')
-rw-r--r--java/src/com/android/inputmethod/keyboard/internal/GestureStroke.java32
-rw-r--r--java/src/com/android/inputmethod/latin/LatinIME.java44
-rw-r--r--java/src/com/android/inputmethod/latin/LatinImeLogger.java2
-rw-r--r--java/src/com/android/inputmethod/latin/Suggest.java4
-rw-r--r--java/src/com/android/inputmethod/latin/UserHistoryDictIOUtils.java8
-rw-r--r--java/src/com/android/inputmethod/latin/Utils.java16
-rw-r--r--java/src/com/android/inputmethod/latin/makedict/BinaryDictIOUtils.java127
-rw-r--r--java/src/com/android/inputmethod/latin/makedict/BinaryDictInputOutput.java195
8 files changed, 247 insertions, 181 deletions
diff --git a/java/src/com/android/inputmethod/keyboard/internal/GestureStroke.java b/java/src/com/android/inputmethod/keyboard/internal/GestureStroke.java
index 825134468..7a8c2409c 100644
--- a/java/src/com/android/inputmethod/keyboard/internal/GestureStroke.java
+++ b/java/src/com/android/inputmethod/keyboard/internal/GestureStroke.java
@@ -25,7 +25,6 @@ public class GestureStroke {
private final ResizableIntArray mXCoordinates = new ResizableIntArray(DEFAULT_CAPACITY);
private final ResizableIntArray mYCoordinates = new ResizableIntArray(DEFAULT_CAPACITY);
private float mLength;
- private float mAngle;
private int mIncrementalRecognitionSize;
private int mLastIncrementalBatchSize;
private long mLastPointTime;
@@ -40,9 +39,6 @@ public class GestureStroke {
private static final int MIN_GESTURE_DURATION = 100; // msec
private static final float MIN_GESTURE_SAMPLING_RATIO_TO_KEY_WIDTH = 1.0f / 6.0f;
private static final float GESTURE_RECOG_SPEED_THRESHOLD = 0.4f; // dip/msec
- private static final float GESTURE_RECOG_CURVATURE_THRESHOLD = (float)(Math.PI / 4.0f);
-
- private static final float DOUBLE_PI = (float)(2.0f * Math.PI);
public GestureStroke(final int pointerId) {
mPointerId = pointerId;
@@ -62,7 +58,6 @@ public class GestureStroke {
public void reset() {
mLength = 0;
- mAngle = 0;
mIncrementalRecognitionSize = 0;
mLastIncrementalBatchSize = 0;
mLastPointTime = 0;
@@ -97,16 +92,6 @@ public class GestureStroke {
mXCoordinates.add(x);
mYCoordinates.add(y);
mLength += dist;
- final float angle = getAngle(lastX, lastY, x, y);
- if (size > 1) {
- final float curvature = getAngleDiff(angle, mAngle);
- if (curvature > GESTURE_RECOG_CURVATURE_THRESHOLD) {
- if (size > mIncrementalRecognitionSize) {
- mIncrementalRecognitionSize = size;
- }
- }
- }
- mAngle = angle;
}
if (!isHistorical) {
@@ -146,21 +131,4 @@ public class GestureStroke {
// java.lang.Math due to the way the JIT optimizes java.lang.Math.
return (float)Math.sqrt(dx * dx + dy * dy);
}
-
- private static float getAngle(final int x1, final int y1, final int x2, final int y2) {
- final int dx = x1 - x2;
- final int dy = y1 - y2;
- if (dx == 0 && dy == 0) return 0;
- // Would it be faster to call atan2f() directly via JNI? Not sure about what the JIT
- // does with Math.atan2().
- return (float)Math.atan2(dy, dx);
- }
-
- private static float getAngleDiff(final float a1, final float a2) {
- final float diff = Math.abs(a1 - a2);
- if (diff > Math.PI) {
- return DOUBLE_PI - diff;
- }
- return diff;
- }
}
diff --git a/java/src/com/android/inputmethod/latin/LatinIME.java b/java/src/com/android/inputmethod/latin/LatinIME.java
index 23334d6f3..bf64b4f08 100644
--- a/java/src/com/android/inputmethod/latin/LatinIME.java
+++ b/java/src/com/android/inputmethod/latin/LatinIME.java
@@ -72,6 +72,7 @@ import com.android.inputmethod.keyboard.KeyboardSwitcher;
import com.android.inputmethod.keyboard.KeyboardView;
import com.android.inputmethod.keyboard.MainKeyboardView;
import com.android.inputmethod.latin.LocaleUtils.RunInLocale;
+import com.android.inputmethod.latin.Utils.Stats;
import com.android.inputmethod.latin.define.ProductionFlag;
import com.android.inputmethod.latin.suggestions.SuggestionStripView;
import com.android.inputmethod.research.ResearchLogger;
@@ -148,7 +149,7 @@ public class LatinIME extends InputMethodService implements KeyboardActionListen
private boolean mIsUserDictionaryAvailable;
private LastComposedWord mLastComposedWord = LastComposedWord.NOT_A_COMPOSED_WORD;
- private WordComposer mWordComposer = new WordComposer();
+ private final WordComposer mWordComposer = new WordComposer();
private RichInputConnection mConnection = new RichInputConnection(this);
// Keep track of the last selection range to decide if we need to show word alternatives
@@ -1331,6 +1332,12 @@ public class LatinIME extends InputMethodService implements KeyboardActionListen
didAutoCorrect = handleSeparator(primaryCode, x, y, spaceState);
} else {
if (SPACE_STATE_PHANTOM == spaceState) {
+ if (ProductionFlag.IS_INTERNAL) {
+ if (mWordComposer.isComposingWord() && mWordComposer.isBatchMode()) {
+ Stats.onAutoCorrection(
+ "", mWordComposer.getTypedWord(), " ", mWordComposer);
+ }
+ }
commitTyped(LastComposedWord.NOT_A_SEPARATOR);
}
final int keyX, keyY;
@@ -1389,6 +1396,11 @@ public class LatinIME extends InputMethodService implements KeyboardActionListen
public void onStartBatchInput() {
mConnection.beginBatchEdit();
if (mWordComposer.isComposingWord()) {
+ if (ProductionFlag.IS_INTERNAL) {
+ if (mWordComposer.isBatchMode()) {
+ Stats.onAutoCorrection("", mWordComposer.getTypedWord(), " ", mWordComposer);
+ }
+ }
commitTyped(LastComposedWord.NOT_A_SEPARATOR);
mExpectingUpdateSelection = true;
// The following is necessary for the case where the user typed something but didn't
@@ -1547,7 +1559,9 @@ public class LatinIME extends InputMethodService implements KeyboardActionListen
}
} else {
if (mLastComposedWord.canRevertCommit()) {
- Utils.Stats.onAutoCorrectionCancellation();
+ if (ProductionFlag.IS_INTERNAL) {
+ Stats.onAutoCorrectionCancellation();
+ }
revertCommit();
return;
}
@@ -1696,7 +1710,9 @@ public class LatinIME extends InputMethodService implements KeyboardActionListen
if (null != mSuggestionStripView) mSuggestionStripView.dismissAddToDictionaryHint();
}
mHandler.postUpdateSuggestionStrip();
- Utils.Stats.onNonSeparator((char)primaryCode, x, y);
+ if (ProductionFlag.IS_INTERNAL) {
+ Utils.Stats.onNonSeparator((char)primaryCode, x, y);
+ }
}
// Returns true if we did an autocorrection, false otherwise.
@@ -1760,8 +1776,9 @@ public class LatinIME extends InputMethodService implements KeyboardActionListen
// already displayed or not, so it's okay.
setPunctuationSuggestions();
}
-
- Utils.Stats.onSeparator((char)primaryCode, x, y);
+ if (ProductionFlag.IS_INTERNAL) {
+ Utils.Stats.onSeparator((char)primaryCode, x, y);
+ }
mHandler.postUpdateShiftState();
return didAutoCorrect;
@@ -1930,7 +1947,10 @@ public class LatinIME extends InputMethodService implements KeyboardActionListen
throw new RuntimeException("We have an auto-correction but the typed word "
+ "is empty? Impossible! I must commit suicide.");
}
- Utils.Stats.onAutoCorrection(typedWord, autoCorrection.toString(), separatorString);
+ if (ProductionFlag.IS_INTERNAL) {
+ Stats.onAutoCorrection(
+ typedWord, autoCorrection.toString(), separatorString, mWordComposer);
+ }
mExpectingUpdateSelection = true;
commitChosenWord(autoCorrection, LastComposedWord.COMMIT_TYPE_DECIDED_WORD,
separatorString);
@@ -2020,8 +2040,10 @@ public class LatinIME extends InputMethodService implements KeyboardActionListen
// If the suggestion is not in the dictionary, the hint should be shown.
&& !AutoCorrection.isValidWord(mSuggest.getUnigramDictionaries(), suggestion, true);
- Utils.Stats.onSeparator((char)Keyboard.CODE_SPACE,
- Constants.NOT_A_COORDINATE, Constants.NOT_A_COORDINATE);
+ if (ProductionFlag.IS_INTERNAL) {
+ Stats.onSeparator((char)Keyboard.CODE_SPACE,
+ Constants.NOT_A_COORDINATE, Constants.NOT_A_COORDINATE);
+ }
if (showingAddToDictionaryHint && mIsUserDictionaryAvailable) {
mSuggestionStripView.showAddToDictionaryHint(
suggestion, mCurrentSettings.mHintToSaveText);
@@ -2138,8 +2160,10 @@ public class LatinIME extends InputMethodService implements KeyboardActionListen
previousWord.toString(), committedWord.toString());
}
mConnection.commitText(originallyTypedWord + mLastComposedWord.mSeparatorString, 1);
- Utils.Stats.onSeparator(mLastComposedWord.mSeparatorString,
- Constants.NOT_A_COORDINATE, Constants.NOT_A_COORDINATE);
+ if (ProductionFlag.IS_INTERNAL) {
+ Stats.onSeparator(mLastComposedWord.mSeparatorString,
+ Constants.NOT_A_COORDINATE, Constants.NOT_A_COORDINATE);
+ }
if (ProductionFlag.IS_EXPERIMENTAL) {
ResearchLogger.latinIME_revertCommit(originallyTypedWord);
}
diff --git a/java/src/com/android/inputmethod/latin/LatinImeLogger.java b/java/src/com/android/inputmethod/latin/LatinImeLogger.java
index a5b4c68d0..9eab19c49 100644
--- a/java/src/com/android/inputmethod/latin/LatinImeLogger.java
+++ b/java/src/com/android/inputmethod/latin/LatinImeLogger.java
@@ -49,7 +49,7 @@ public class LatinImeLogger implements SharedPreferences.OnSharedPreferenceChang
}
public static void logOnAutoCorrectionForGeometric(String before, String after,
- int separatorCode, int[] xCoordinates, int[] yCoordinates, int[] relativeTimes) {
+ int separatorCode, InputPointers inputPointers) {
}
public static void logOnAutoCorrectionCancelled() {
diff --git a/java/src/com/android/inputmethod/latin/Suggest.java b/java/src/com/android/inputmethod/latin/Suggest.java
index 0418d3166..278c4b9ce 100644
--- a/java/src/com/android/inputmethod/latin/Suggest.java
+++ b/java/src/com/android/inputmethod/latin/Suggest.java
@@ -306,6 +306,10 @@ public class Suggest {
wordComposer, prevWordForBigram, proximityInfo, sessionId));
}
+ for (SuggestedWordInfo wordInfo : suggestionsSet) {
+ LatinImeLogger.onAddSuggestedWord(wordInfo.mWord.toString(), wordInfo.mSourceDict);
+ }
+
final ArrayList<SuggestedWordInfo> suggestionsContainer =
CollectionUtils.newArrayList(suggestionsSet);
final int suggestionsCount = suggestionsContainer.size();
diff --git a/java/src/com/android/inputmethod/latin/UserHistoryDictIOUtils.java b/java/src/com/android/inputmethod/latin/UserHistoryDictIOUtils.java
index 550e4e58b..4a3d11aa1 100644
--- a/java/src/com/android/inputmethod/latin/UserHistoryDictIOUtils.java
+++ b/java/src/com/android/inputmethod/latin/UserHistoryDictIOUtils.java
@@ -18,6 +18,7 @@ package com.android.inputmethod.latin;
import android.util.Log;
+import com.android.inputmethod.latin.makedict.BinaryDictIOUtils;
import com.android.inputmethod.latin.makedict.BinaryDictInputOutput;
import com.android.inputmethod.latin.makedict.BinaryDictInputOutput.FusionDictionaryBufferInterface;
import com.android.inputmethod.latin.makedict.FormatSpec.FormatOptions;
@@ -96,6 +97,11 @@ public class UserHistoryDictIOUtils {
public void put(final byte b) {
mBuffer[mPosition++] = b;
}
+
+ @Override
+ public int limit() {
+ return mBuffer.length;
+ }
}
/**
@@ -162,7 +168,7 @@ public class UserHistoryDictIOUtils {
final Map<Integer, ArrayList<PendingAttribute>> bigrams = CollectionUtils.newTreeMap();
try {
- BinaryDictInputOutput.readUnigramsAndBigramsBinary(buffer, unigrams, frequencies,
+ BinaryDictIOUtils.readUnigramsAndBigramsBinary(buffer, unigrams, frequencies,
bigrams);
addWordsFromWordMap(unigrams, frequencies, bigrams, dict);
} catch (IOException e) {
diff --git a/java/src/com/android/inputmethod/latin/Utils.java b/java/src/com/android/inputmethod/latin/Utils.java
index 63b642821..876bc8e79 100644
--- a/java/src/com/android/inputmethod/latin/Utils.java
+++ b/java/src/com/android/inputmethod/latin/Utils.java
@@ -412,14 +412,24 @@ public final class Utils {
}
public static void onAutoCorrection(final String typedWord, final String correctedWord,
- final String separatorString) {
- if (TextUtils.isEmpty(typedWord)) return;
+ final String separatorString, final WordComposer wordComposer) {
+ final boolean isBatchMode = wordComposer.isBatchMode();
+ if (!isBatchMode && TextUtils.isEmpty(typedWord)) return;
// TODO: this fails when the separator is more than 1 code point long, but
// the backend can't handle it yet. The only case when this happens is with
// smileys and other multi-character keys.
final int codePoint = TextUtils.isEmpty(separatorString) ? Constants.NOT_A_CODE
: separatorString.codePointAt(0);
- LatinImeLogger.logOnAutoCorrectionForTyping(typedWord, correctedWord, codePoint);
+ if (!isBatchMode) {
+ LatinImeLogger.logOnAutoCorrectionForTyping(typedWord, correctedWord, codePoint);
+ } else {
+ if (!TextUtils.isEmpty(correctedWord)) {
+ // We must make sure that InputPointer contains only the relative timestamps,
+ // not actual timestamps.
+ LatinImeLogger.logOnAutoCorrectionForGeometric(
+ "", correctedWord, codePoint, wordComposer.getInputPointers());
+ }
+ }
}
public static void onAutoCorrectionCancellation() {
diff --git a/java/src/com/android/inputmethod/latin/makedict/BinaryDictIOUtils.java b/java/src/com/android/inputmethod/latin/makedict/BinaryDictIOUtils.java
new file mode 100644
index 000000000..1a85e71ce
--- /dev/null
+++ b/java/src/com/android/inputmethod/latin/makedict/BinaryDictIOUtils.java
@@ -0,0 +1,127 @@
+/*
+ * Copyright (C) 2012 The Android Open Source Project
+ *
+ * Licensed under the Apache License, Version 2.0 (the "License"); you may not
+ * use this file except in compliance with the License. You may obtain a copy of
+ * the License at
+ *
+ * http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS, WITHOUT
+ * WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the
+ * License for the specific language governing permissions and limitations under
+ * the License.
+ */
+
+package com.android.inputmethod.latin.makedict;
+
+import com.android.inputmethod.latin.makedict.BinaryDictInputOutput;
+import com.android.inputmethod.latin.makedict.BinaryDictInputOutput.FusionDictionaryBufferInterface;
+import com.android.inputmethod.latin.makedict.FormatSpec.FileHeader;
+import com.android.inputmethod.latin.makedict.FormatSpec.FormatOptions;
+
+import java.io.IOException;
+import java.util.ArrayList;
+import java.util.Map;
+import java.util.Stack;
+
+public class BinaryDictIOUtils {
+ private static final boolean DBG = false;
+
+ private static class Position {
+ public static final int NOT_READ_GROUPCOUNT = -1;
+
+ public int mAddress;
+ public int mNumOfCharGroup;
+ public int mPosition;
+ public int mLength;
+
+ public Position(int address, int length) {
+ mAddress = address;
+ mLength = length;
+ mNumOfCharGroup = NOT_READ_GROUPCOUNT;
+ }
+ }
+
+ /**
+ * Tours all node without recursive call.
+ */
+ private static void readUnigramsAndBigramsBinaryInner(
+ final FusionDictionaryBufferInterface buffer, final int headerSize,
+ final Map<Integer, String> words, final Map<Integer, Integer> frequencies,
+ final Map<Integer, ArrayList<PendingAttribute>> bigrams,
+ final FormatOptions formatOptions) {
+ int[] pushedChars = new int[FormatSpec.MAX_WORD_LENGTH + 1];
+
+ Stack<Position> stack = new Stack<Position>();
+ int index = 0;
+
+ Position initPos = new Position(headerSize, 0);
+ stack.push(initPos);
+
+ while (!stack.empty()) {
+ Position p = stack.peek();
+
+ if (DBG) {
+ MakedictLog.d("read: address=" + p.mAddress + ", numOfCharGroup=" +
+ p.mNumOfCharGroup + ", position=" + p.mPosition + ", length=" + p.mLength);
+ }
+
+ if (buffer.position() != p.mAddress) buffer.position(p.mAddress);
+ if (index != p.mLength) index = p.mLength;
+
+ if (p.mNumOfCharGroup == Position.NOT_READ_GROUPCOUNT) {
+ p.mNumOfCharGroup = BinaryDictInputOutput.readCharGroupCount(buffer);
+ p.mAddress += BinaryDictInputOutput.getGroupCountSize(p.mNumOfCharGroup);
+ p.mPosition = 0;
+ }
+
+ CharGroupInfo info = BinaryDictInputOutput.readCharGroup(buffer,
+ p.mAddress - headerSize, formatOptions);
+ for (int i = 0; i < info.mCharacters.length; ++i) {
+ pushedChars[index++] = info.mCharacters[i];
+ }
+ p.mPosition++;
+
+ if (info.mFrequency != FusionDictionary.CharGroup.NOT_A_TERMINAL) { // found word
+ words.put(info.mOriginalAddress, new String(pushedChars, 0, index));
+ frequencies.put(info.mOriginalAddress, info.mFrequency);
+ if (info.mBigrams != null) bigrams.put(info.mOriginalAddress, info.mBigrams);
+ }
+
+ if (p.mPosition == p.mNumOfCharGroup) {
+ stack.pop();
+ } else {
+ // the node has more groups.
+ p.mAddress = buffer.position();
+ }
+
+ if (BinaryDictInputOutput.hasChildrenAddress(info.mChildrenAddress)) {
+ Position childrenPos = new Position(info.mChildrenAddress + headerSize, index);
+ stack.push(childrenPos);
+ }
+ }
+ }
+
+ /**
+ * Reads unigrams and bigrams from the binary file.
+ * Doesn't make the memory representation of the dictionary.
+ *
+ * @param buffer the buffer to read.
+ * @param words the map to store the address as a key and the word as a value.
+ * @param frequencies the map to store the address as a key and the frequency as a value.
+ * @param bigrams the map to store the address as a key and the list of address as a value.
+ * @throws IOException
+ * @throws UnsupportedFormatException
+ */
+ public static void readUnigramsAndBigramsBinary(final FusionDictionaryBufferInterface buffer,
+ final Map<Integer, String> words, final Map<Integer, Integer> frequencies,
+ final Map<Integer, ArrayList<PendingAttribute>> bigrams) throws IOException,
+ UnsupportedFormatException {
+ // Read header
+ final FileHeader header = BinaryDictInputOutput.readHeader(buffer);
+ readUnigramsAndBigramsBinaryInner(buffer, header.mHeaderSize, words, frequencies, bigrams,
+ header.mFormatOptions);
+ }
+}
diff --git a/java/src/com/android/inputmethod/latin/makedict/BinaryDictInputOutput.java b/java/src/com/android/inputmethod/latin/makedict/BinaryDictInputOutput.java
index 72d12299b..c865702d6 100644
--- a/java/src/com/android/inputmethod/latin/makedict/BinaryDictInputOutput.java
+++ b/java/src/com/android/inputmethod/latin/makedict/BinaryDictInputOutput.java
@@ -63,6 +63,7 @@ public class BinaryDictInputOutput {
public int position();
public void position(int newPosition);
public void put(final byte b);
+ public int limit();
}
public static final class ByteBufferWrapper implements FusionDictionaryBufferInterface {
@@ -107,6 +108,11 @@ public class BinaryDictInputOutput {
public void put(final byte b) {
mBuffer.put(b);
}
+
+ @Override
+ public int limit() {
+ return mBuffer.limit();
+ }
}
/**
@@ -284,7 +290,7 @@ public class BinaryDictInputOutput {
* @param count the group count
* @return the size of the group count, either 1 or 2 bytes.
*/
- private static int getGroupCountSize(final int count) {
+ public static int getGroupCountSize(final int count) {
if (FormatSpec.MAX_CHARGROUPS_FOR_ONE_BYTE_CHARGROUP_COUNT >= count) {
return 1;
} else if (FormatSpec.MAX_CHARGROUPS_IN_A_NODE >= count) {
@@ -379,7 +385,7 @@ public class BinaryDictInputOutput {
/**
* Helper method to hide the actual value of the no children address.
*/
- private static boolean hasChildrenAddress(final int address) {
+ public static boolean hasChildrenAddress(final int address) {
return FormatSpec.NO_CHILDREN_ADDRESS != address;
}
@@ -1099,7 +1105,7 @@ public class BinaryDictInputOutput {
// readDictionaryBinary is the public entry point for them.
private static final int[] CHARACTER_BUFFER = new int[FormatSpec.MAX_WORD_LENGTH];
- private static CharGroupInfo readCharGroup(final FusionDictionaryBufferInterface buffer,
+ public static CharGroupInfo readCharGroup(final FusionDictionaryBufferInterface buffer,
final int originalGroupAddress, final FormatOptions options) {
int addressPointer = originalGroupAddress;
final int flags = buffer.readUnsignedByte();
@@ -1212,7 +1218,7 @@ public class BinaryDictInputOutput {
/**
* Reads and returns the char group count out of a buffer and forwards the pointer.
*/
- private static int readCharGroupCount(final FusionDictionaryBufferInterface buffer) {
+ public static int readCharGroupCount(final FusionDictionaryBufferInterface buffer) {
final int msb = buffer.readUnsignedByte();
if (FormatSpec.MAX_CHARGROUPS_FOR_ONE_BYTE_CHARGROUP_COUNT >= msb) {
return msb;
@@ -1341,146 +1347,67 @@ public class BinaryDictInputOutput {
final Map<Integer, Node> reverseNodeMap, final Map<Integer, CharGroup> reverseGroupMap,
final FormatOptions options)
throws IOException {
- final int nodeOrigin = buffer.position() - headerSize;
- final int count = readCharGroupCount(buffer);
final ArrayList<CharGroup> nodeContents = new ArrayList<CharGroup>();
- int groupOffset = nodeOrigin + getGroupCountSize(count);
- for (int i = count; i > 0; --i) {
- CharGroupInfo info = readCharGroup(buffer, groupOffset, options);
- ArrayList<WeightedString> shortcutTargets = info.mShortcutTargets;
- ArrayList<WeightedString> bigrams = null;
- if (null != info.mBigrams) {
- bigrams = new ArrayList<WeightedString>();
- for (PendingAttribute bigram : info.mBigrams) {
- final String word = getWordAtAddress(
- buffer, headerSize, bigram.mAddress, options);
- bigrams.add(new WeightedString(word, bigram.mFrequency));
+ final int nodeOrigin = buffer.position() - headerSize;
+
+ do { // Scan the linked-list node.
+ final int nodeHeadPosition = buffer.position() - headerSize;
+ final int count = readCharGroupCount(buffer);
+ int groupOffset = nodeHeadPosition + getGroupCountSize(count);
+ for (int i = count; i > 0; --i) { // Scan the array of CharGroup.
+ CharGroupInfo info = readCharGroup(buffer, groupOffset, options);
+ ArrayList<WeightedString> shortcutTargets = info.mShortcutTargets;
+ ArrayList<WeightedString> bigrams = null;
+ if (null != info.mBigrams) {
+ bigrams = new ArrayList<WeightedString>();
+ for (PendingAttribute bigram : info.mBigrams) {
+ final String word = getWordAtAddress(
+ buffer, headerSize, bigram.mAddress, options);
+ bigrams.add(new WeightedString(word, bigram.mFrequency));
+ }
+ }
+ if (hasChildrenAddress(info.mChildrenAddress)) {
+ Node children = reverseNodeMap.get(info.mChildrenAddress);
+ if (null == children) {
+ final int currentPosition = buffer.position();
+ buffer.position(info.mChildrenAddress + headerSize);
+ children = readNode(
+ buffer, headerSize, reverseNodeMap, reverseGroupMap, options);
+ buffer.position(currentPosition);
+ }
+ nodeContents.add(
+ new CharGroup(info.mCharacters, shortcutTargets, bigrams,
+ info.mFrequency,
+ 0 != (info.mFlags & FormatSpec.FLAG_IS_NOT_A_WORD),
+ 0 != (info.mFlags & FormatSpec.FLAG_IS_BLACKLISTED), children));
+ } else {
+ nodeContents.add(
+ new CharGroup(info.mCharacters, shortcutTargets, bigrams,
+ info.mFrequency,
+ 0 != (info.mFlags & FormatSpec.FLAG_IS_NOT_A_WORD),
+ 0 != (info.mFlags & FormatSpec.FLAG_IS_BLACKLISTED)));
}
+ groupOffset = info.mEndAddress;
}
- if (hasChildrenAddress(info.mChildrenAddress)) {
- Node children = reverseNodeMap.get(info.mChildrenAddress);
- if (null == children) {
- final int currentPosition = buffer.position();
- buffer.position(info.mChildrenAddress + headerSize);
- children = readNode(
- buffer, headerSize, reverseNodeMap, reverseGroupMap, options);
- buffer.position(currentPosition);
+
+ // reach the end of the array.
+ if (options.mHasLinkedListNode) {
+ final int nextAddress = buffer.readUnsignedInt24();
+ if (nextAddress >= 0 && nextAddress < buffer.limit()) {
+ buffer.position(nextAddress);
+ } else {
+ break;
}
- nodeContents.add(
- new CharGroup(info.mCharacters, shortcutTargets, bigrams, info.mFrequency,
- 0 != (info.mFlags & FormatSpec.FLAG_IS_NOT_A_WORD),
- 0 != (info.mFlags & FormatSpec.FLAG_IS_BLACKLISTED), children));
- } else {
- nodeContents.add(
- new CharGroup(info.mCharacters, shortcutTargets, bigrams, info.mFrequency,
- 0 != (info.mFlags & FormatSpec.FLAG_IS_NOT_A_WORD),
- 0 != (info.mFlags & FormatSpec.FLAG_IS_BLACKLISTED)));
}
- groupOffset = info.mEndAddress;
- }
+ } while (options.mHasLinkedListNode &&
+ buffer.position() != FormatSpec.NO_FORWARD_LINK_ADDRESS);
+
final Node node = new Node(nodeContents);
node.mCachedAddress = nodeOrigin;
reverseNodeMap.put(node.mCachedAddress, node);
return node;
}
- // TODO: move these methods (readUnigramsAndBigramsBinary(|Inner)) and an inner class (Position)
- // out of this class.
- private static class Position {
- public static final int NOT_READ_GROUPCOUNT = -1;
-
- public int mAddress;
- public int mNumOfCharGroup;
- public int mPosition;
- public int mLength;
-
- public Position(int address, int length) {
- mAddress = address;
- mLength = length;
- mNumOfCharGroup = NOT_READ_GROUPCOUNT;
- }
- }
-
- /**
- * Tours all node without recursive call.
- */
- private static void readUnigramsAndBigramsBinaryInner(
- final FusionDictionaryBufferInterface buffer, final int headerSize,
- final Map<Integer, String> words, final Map<Integer, Integer> frequencies,
- final Map<Integer, ArrayList<PendingAttribute>> bigrams,
- final FormatOptions formatOptions) {
- int[] pushedChars = new int[FormatSpec.MAX_WORD_LENGTH + 1];
-
- Stack<Position> stack = new Stack<Position>();
- int index = 0;
-
- Position initPos = new Position(headerSize, 0);
- stack.push(initPos);
-
- while (!stack.empty()) {
- Position p = stack.peek();
-
- if (DBG) {
- MakedictLog.d("read: address=" + p.mAddress + ", numOfCharGroup=" +
- p.mNumOfCharGroup + ", position=" + p.mPosition + ", length=" + p.mLength);
- }
-
- if (buffer.position() != p.mAddress) buffer.position(p.mAddress);
- if (index != p.mLength) index = p.mLength;
-
- if (p.mNumOfCharGroup == Position.NOT_READ_GROUPCOUNT) {
- p.mNumOfCharGroup = readCharGroupCount(buffer);
- p.mAddress += getGroupCountSize(p.mNumOfCharGroup);
- p.mPosition = 0;
- }
-
- CharGroupInfo info = readCharGroup(buffer, p.mAddress - headerSize, formatOptions);
- for (int i = 0; i < info.mCharacters.length; ++i) {
- pushedChars[index++] = info.mCharacters[i];
- }
- p.mPosition++;
-
- if (info.mFrequency != FusionDictionary.CharGroup.NOT_A_TERMINAL) { // found word
- words.put(info.mOriginalAddress, new String(pushedChars, 0, index));
- frequencies.put(info.mOriginalAddress, info.mFrequency);
- if (info.mBigrams != null) bigrams.put(info.mOriginalAddress, info.mBigrams);
- }
-
- if (p.mPosition == p.mNumOfCharGroup) {
- stack.pop();
- } else {
- // the node has more groups.
- p.mAddress = buffer.position();
- }
-
- if (hasChildrenAddress(info.mChildrenAddress)) {
- Position childrenPos = new Position(info.mChildrenAddress + headerSize, index);
- stack.push(childrenPos);
- }
- }
- }
-
- /**
- * Reads unigrams and bigrams from the binary file.
- * Doesn't make the memory representation of the dictionary.
- *
- * @param buffer the buffer to read.
- * @param words the map to store the address as a key and the word as a value.
- * @param frequencies the map to store the address as a key and the frequency as a value.
- * @param bigrams the map to store the address as a key and the list of address as a value.
- * @throws IOException
- * @throws UnsupportedFormatException
- */
- public static void readUnigramsAndBigramsBinary(final FusionDictionaryBufferInterface buffer,
- final Map<Integer, String> words, final Map<Integer, Integer> frequencies,
- final Map<Integer, ArrayList<PendingAttribute>> bigrams) throws IOException,
- UnsupportedFormatException {
- // Read header
- final FileHeader header = readHeader(buffer);
- readUnigramsAndBigramsBinaryInner(buffer, header.mHeaderSize, words, frequencies, bigrams,
- header.mFormatOptions);
- }
-
/**
* Helper function to get the binary format version from the header.
* @throws IOException
@@ -1517,7 +1444,7 @@ public class BinaryDictInputOutput {
* @throws IOException
* @throws UnsupportedFormatException
*/
- private static FileHeader readHeader(final FusionDictionaryBufferInterface buffer)
+ public static FileHeader readHeader(final FusionDictionaryBufferInterface buffer)
throws IOException, UnsupportedFormatException {
final int version = checkFormatVersion(buffer);
final int optionsFlags = buffer.readUnsignedShort();