BinaryDictionary.java revision 50db5f0fffdc35e9fe0bed27e7d630d571727e06 
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 NativeSuggestOptions mNativeSuggestOptions = new NativeSuggestOptions();
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        mNativeSuggestOptions.setUseFullEditDistance(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            int[] suggestOptions, int[] prevWordCodePointArray,
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            final boolean blockOffensiveWords) {
112        return getSuggestionsWithSessionId(composer, prevWord, proximityInfo, blockOffensiveWords,
113                0 /* sessionId */);
114    }
115
116    @Override
117    public ArrayList<SuggestedWordInfo> getSuggestionsWithSessionId(final WordComposer composer,
118            final String prevWord, final ProximityInfo proximityInfo,
119            final boolean blockOffensiveWords, final int sessionId) {
120        if (!isValidDictionary()) return null;
121
122        Arrays.fill(mInputCodePoints, Constants.NOT_A_CODE);
123        // TODO: toLowerCase in the native code
124        final int[] prevWordCodePointArray = (null == prevWord)
125                ? null : StringUtils.toCodePointArray(prevWord);
126        final int composerSize = composer.size();
127
128        final boolean isGesture = composer.isBatchMode();
129        if (composerSize <= 1 || !isGesture) {
130            if (composerSize > MAX_WORD_LENGTH - 1) return null;
131            for (int i = 0; i < composerSize; i++) {
132                mInputCodePoints[i] = composer.getCodeAt(i);
133            }
134        }
135
136        final InputPointers ips = composer.getInputPointers();
137        final int inputSize = isGesture ? ips.getPointerSize() : composerSize;
138        mNativeSuggestOptions.setIsGesture(isGesture);
139        // proximityInfo and/or prevWordForBigrams may not be null.
140        final int count = getSuggestionsNative(mNativeDict, proximityInfo.getNativeProximityInfo(),
141                getTraverseSession(sessionId).getSession(), ips.getXCoordinates(),
142                ips.getYCoordinates(), ips.getTimes(), ips.getPointerIds(), mInputCodePoints,
143                inputSize, 0 /* commitPoint */, mNativeSuggestOptions.getOptions(),
144                prevWordCodePointArray, mOutputCodePoints, mOutputScores, mSpaceIndices,
145                mOutputTypes);
146        final ArrayList<SuggestedWordInfo> suggestions = CollectionUtils.newArrayList();
147        for (int j = 0; j < count; ++j) {
148            final int start = j * MAX_WORD_LENGTH;
149            int len = 0;
150            while (len < MAX_WORD_LENGTH && mOutputCodePoints[start + len] != 0) {
151                ++len;
152            }
153            if (len > 0) {
154                final int flags = mOutputTypes[j] & SuggestedWordInfo.KIND_MASK_FLAGS;
155                if (blockOffensiveWords
156                        && 0 != (flags & SuggestedWordInfo.KIND_FLAG_POSSIBLY_OFFENSIVE)
157                        && 0 == (flags & SuggestedWordInfo.KIND_FLAG_EXACT_MATCH)) {
158                    // If we block potentially offensive words, and if the word is possibly
159                    // offensive, then we don't output it unless it's also an exact match.
160                    continue;
161                }
162                final int kind = mOutputTypes[j] & SuggestedWordInfo.KIND_MASK_KIND;
163                final int score = SuggestedWordInfo.KIND_WHITELIST == kind
164                        ? SuggestedWordInfo.MAX_SCORE : mOutputScores[j];
165                // TODO: check that all users of the `kind' parameter are ready to accept
166                // flags too and pass mOutputTypes[j] instead of kind
167                suggestions.add(new SuggestedWordInfo(new String(mOutputCodePoints, start, len),
168                        score, kind, mDictType));
169            }
170        }
171        return suggestions;
172    }
173
174    public boolean isValidDictionary() {
175        return mNativeDict != 0;
176    }
177
178    public static float calcNormalizedScore(final String before, final String after,
179            final int score) {
180        return calcNormalizedScoreNative(StringUtils.toCodePointArray(before),
181                StringUtils.toCodePointArray(after), score);
182    }
183
184    public static int editDistance(final String before, final String after) {
185        if (before == null || after == null) {
186            throw new IllegalArgumentException();
187        }
188        return editDistanceNative(StringUtils.toCodePointArray(before),
189                StringUtils.toCodePointArray(after));
190    }
191
192    @Override
193    public boolean isValidWord(final String word) {
194        return getFrequency(word) >= 0;
195    }
196
197    @Override
198    public int getFrequency(final String word) {
199        if (word == null) return -1;
200        int[] codePoints = StringUtils.toCodePointArray(word);
201        return getProbabilityNative(mNativeDict, codePoints);
202    }
203
204    // TODO: Add a batch process version (isValidBigramMultiple?) to avoid excessive numbers of jni
205    // calls when checking for changes in an entire dictionary.
206    public boolean isValidBigram(final String word1, final String word2) {
207        if (TextUtils.isEmpty(word1) || TextUtils.isEmpty(word2)) return false;
208        final int[] codePoints1 = StringUtils.toCodePointArray(word1);
209        final int[] codePoints2 = StringUtils.toCodePointArray(word2);
210        return isValidBigramNative(mNativeDict, codePoints1, codePoints2);
211    }
212
213    @Override
214    public void close() {
215        synchronized (mDicTraverseSessions) {
216            final int sessionsSize = mDicTraverseSessions.size();
217            for (int index = 0; index < sessionsSize; ++index) {
218                final DicTraverseSession traverseSession = mDicTraverseSessions.valueAt(index);
219                if (traverseSession != null) {
220                    traverseSession.close();
221                }
222            }
223        }
224        closeInternal();
225    }
226
227    private synchronized void closeInternal() {
228        if (mNativeDict != 0) {
229            closeNative(mNativeDict);
230            mNativeDict = 0;
231        }
232    }
233
234    @Override
235    protected void finalize() throws Throwable {
236        try {
237            closeInternal();
238        } finally {
239            super.finalize();
240        }
241    }
242}
243