BinaryDictionary.java revision 99e998286d71cf698d0a809a29b15d1a231ebbb1 
1/*
2 * Copyright (C) 2008 The Android Open Source Project
3 *
4 * Licensed under the Apache License, Version 2.0 (the "License");
5 * you may not use this file except in compliance with the License.
6 * You may obtain a copy of the License at
7 *
8 *      http://www.apache.org/licenses/LICENSE-2.0
9 *
10 * Unless required by applicable law or agreed to in writing, software
11 * distributed under the License is distributed on an "AS IS" BASIS,
12 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
13 * See the License for the specific language governing permissions and
14 * limitations under the License.
15 */
16
17package com.android.inputmethod.latin;
18
19import android.text.TextUtils;
20import android.util.SparseArray;
21
22import com.android.inputmethod.keyboard.ProximityInfo;
23import com.android.inputmethod.latin.SuggestedWords.SuggestedWordInfo;
24
25import java.util.ArrayList;
26import java.util.Arrays;
27import java.util.Locale;
28
29/**
30 * Implements a static, compacted, binary dictionary of standard words.
31 */
32public final class BinaryDictionary extends Dictionary {
33    private static final String TAG = BinaryDictionary.class.getSimpleName();
34
35    // Must be equal to MAX_WORD_LENGTH in native/jni/src/defines.h
36    private static final int MAX_WORD_LENGTH = Constants.Dictionary.MAX_WORD_LENGTH;
37    // Must be equal to MAX_RESULTS in native/jni/src/defines.h
38    private static final int MAX_RESULTS = 18;
39
40    private long mNativeDict;
41    private final Locale mLocale;
42    private final int[] mInputCodePoints = new int[MAX_WORD_LENGTH];
43    private final int[] mOutputCodePoints = new int[MAX_WORD_LENGTH * MAX_RESULTS];
44    private final int[] mSpaceIndices = new int[MAX_RESULTS];
45    private final int[] mOutputScores = new int[MAX_RESULTS];
46    private final int[] mOutputTypes = new int[MAX_RESULTS];
47
48    private final boolean mUseFullEditDistance;
49
50    private final SparseArray<DicTraverseSession> mDicTraverseSessions =
51            CollectionUtils.newSparseArray();
52
53    // TODO: There should be a way to remove used DicTraverseSession objects from
54    // {@code mDicTraverseSessions}.
55    private DicTraverseSession getTraverseSession(final int traverseSessionId) {
56        synchronized(mDicTraverseSessions) {
57            DicTraverseSession traverseSession = mDicTraverseSessions.get(traverseSessionId);
58            if (traverseSession == null) {
59                traverseSession = mDicTraverseSessions.get(traverseSessionId);
60                if (traverseSession == null) {
61                    traverseSession = new DicTraverseSession(mLocale, mNativeDict);
62                    mDicTraverseSessions.put(traverseSessionId, traverseSession);
63                }
64            }
65            return traverseSession;
66        }
67    }
68
69    /**
70     * Constructor for the binary dictionary. This is supposed to be called from the
71     * dictionary factory.
72     * @param filename the name of the file to read through native code.
73     * @param offset the offset of the dictionary data within the file.
74     * @param length the length of the binary data.
75     * @param useFullEditDistance whether to use the full edit distance in suggestions
76     * @param dictType the dictionary type, as a human-readable string
77     */
78    public BinaryDictionary(final String filename, final long offset, final long length,
79            final boolean useFullEditDistance, final Locale locale, final String dictType) {
80        super(dictType);
81        mLocale = locale;
82        mUseFullEditDistance = useFullEditDistance;
83        loadDictionary(filename, offset, length);
84    }
85
86    static {
87        JniUtils.loadNativeLibrary();
88    }
89
90    private static native long openNative(String sourceDir, long dictOffset, long dictSize);
91    private static native void closeNative(long dict);
92    private static native int getProbabilityNative(long dict, int[] word);
93    private static native boolean isValidBigramNative(long dict, int[] word1, int[] word2);
94    private static native int getSuggestionsNative(long dict, long proximityInfo,
95            long traverseSession, int[] xCoordinates, int[] yCoordinates, int[] times,
96            int[] pointerIds, int[] inputCodePoints, int inputSize, int commitPoint,
97            boolean isGesture, int[] prevWordCodePointArray, boolean useFullEditDistance,
98            int[] outputCodePoints, int[] outputScores, int[] outputIndices, int[] outputTypes);
99    private static native float calcNormalizedScoreNative(int[] before, int[] after, int score);
100    private static native int editDistanceNative(int[] before, int[] after);
101
102    // TODO: Move native dict into session
103    private final void loadDictionary(final String path, final long startOffset,
104            final long length) {
105        mNativeDict = openNative(path, startOffset, length);
106    }
107
108    @Override
109    public ArrayList<SuggestedWordInfo> getSuggestions(final WordComposer composer,
110            final String prevWord, final ProximityInfo proximityInfo) {
111        return getSuggestionsWithSessionId(composer, prevWord, proximityInfo, 0);
112    }
113
114    @Override
115    public ArrayList<SuggestedWordInfo> getSuggestionsWithSessionId(final WordComposer composer,
116            final String prevWord, final ProximityInfo proximityInfo, int sessionId) {
117        if (!isValidDictionary()) return null;
118
119        Arrays.fill(mInputCodePoints, Constants.NOT_A_CODE);
120        // TODO: toLowerCase in the native code
121        final int[] prevWordCodePointArray = (null == prevWord)
122                ? null : StringUtils.toCodePointArray(prevWord);
123        final int composerSize = composer.size();
124
125        final boolean isGesture = composer.isBatchMode();
126        if (composerSize <= 1 || !isGesture) {
127            if (composerSize > MAX_WORD_LENGTH - 1) return null;
128            for (int i = 0; i < composerSize; i++) {
129                mInputCodePoints[i] = composer.getCodeAt(i);
130            }
131        }
132
133        final InputPointers ips = composer.getInputPointers();
134        final int inputSize = isGesture ? ips.getPointerSize() : composerSize;
135        // proximityInfo and/or prevWordForBigrams may not be null.
136        final int count = getSuggestionsNative(mNativeDict, proximityInfo.getNativeProximityInfo(),
137                getTraverseSession(sessionId).getSession(), ips.getXCoordinates(),
138                ips.getYCoordinates(), ips.getTimes(), ips.getPointerIds(), mInputCodePoints,
139                inputSize, 0 /* commitPoint */, isGesture, prevWordCodePointArray,
140                mUseFullEditDistance, mOutputCodePoints, mOutputScores, mSpaceIndices,
141                mOutputTypes);
142        final ArrayList<SuggestedWordInfo> suggestions = CollectionUtils.newArrayList();
143        for (int j = 0; j < count; ++j) {
144            final int start = j * MAX_WORD_LENGTH;
145            int len = 0;
146            while (len < MAX_WORD_LENGTH && mOutputCodePoints[start + len] != 0) {
147                ++len;
148            }
149            if (len > 0) {
150                final int flags = mOutputTypes[j] & SuggestedWordInfo.KIND_MASK_FLAGS;
151                if (0 != (flags & SuggestedWordInfo.KIND_FLAG_POSSIBLY_OFFENSIVE)
152                        && 0 == (flags & SuggestedWordInfo.KIND_FLAG_EXACT_MATCH)) {
153                    // If the word is possibly offensive, we don't output it unless it's also
154                    // an exact match.
155                    continue;
156                }
157                final int kind = mOutputTypes[j] & SuggestedWordInfo.KIND_MASK_KIND;
158                final int score = SuggestedWordInfo.KIND_WHITELIST == kind
159                        ? SuggestedWordInfo.MAX_SCORE : mOutputScores[j];
160                // TODO: check that all users of the `kind' parameter are ready to accept
161                // flags too and pass mOutputTypes[j] instead of kind
162                suggestions.add(new SuggestedWordInfo(new String(mOutputCodePoints, start, len),
163                        score, kind, mDictType));
164            }
165        }
166        return suggestions;
167    }
168
169    public boolean isValidDictionary() {
170        return mNativeDict != 0;
171    }
172
173    public static float calcNormalizedScore(final String before, final String after,
174            final int score) {
175        return calcNormalizedScoreNative(StringUtils.toCodePointArray(before),
176                StringUtils.toCodePointArray(after), score);
177    }
178
179    public static int editDistance(final String before, final String after) {
180        if (before == null || after == null) {
181            throw new IllegalArgumentException();
182        }
183        return editDistanceNative(StringUtils.toCodePointArray(before),
184                StringUtils.toCodePointArray(after));
185    }
186
187    @Override
188    public boolean isValidWord(final String word) {
189        return getFrequency(word) >= 0;
190    }
191
192    @Override
193    public int getFrequency(final String word) {
194        if (word == null) return -1;
195        int[] codePoints = StringUtils.toCodePointArray(word);
196        return getProbabilityNative(mNativeDict, codePoints);
197    }
198
199    // TODO: Add a batch process version (isValidBigramMultiple?) to avoid excessive numbers of jni
200    // calls when checking for changes in an entire dictionary.
201    public boolean isValidBigram(final String word1, final String word2) {
202        if (TextUtils.isEmpty(word1) || TextUtils.isEmpty(word2)) return false;
203        final int[] codePoints1 = StringUtils.toCodePointArray(word1);
204        final int[] codePoints2 = StringUtils.toCodePointArray(word2);
205        return isValidBigramNative(mNativeDict, codePoints1, codePoints2);
206    }
207
208    @Override
209    public void close() {
210        synchronized (mDicTraverseSessions) {
211            final int sessionsSize = mDicTraverseSessions.size();
212            for (int index = 0; index < sessionsSize; ++index) {
213                final DicTraverseSession traverseSession = mDicTraverseSessions.valueAt(index);
214                if (traverseSession != null) {
215                    traverseSession.close();
216                }
217            }
218        }
219        closeInternal();
220    }
221
222    private synchronized void closeInternal() {
223        if (mNativeDict != 0) {
224            closeNative(mNativeDict);
225            mNativeDict = 0;
226        }
227    }
228
229    @Override
230    protected void finalize() throws Throwable {
231        try {
232            closeInternal();
233        } finally {
234            super.finalize();
235        }
236    }
237}
238