dictionary.cpp revision 3e75c59133000d16d3b5606f67d4ec60988851d8 
1/*
2 * Copyright (C) 2009, 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
17#define LOG_TAG "LatinIME: dictionary.cpp"
18
19#include "suggest/core/dictionary/dictionary.h"
20
21#include "defines.h"
22#include "suggest/core/dictionary/dictionary_utils.h"
23#include "suggest/core/policy/dictionary_header_structure_policy.h"
24#include "suggest/core/result/suggestion_results.h"
25#include "suggest/core/session/dic_traverse_session.h"
26#include "suggest/core/session/prev_words_info.h"
27#include "suggest/core/suggest.h"
28#include "suggest/core/suggest_options.h"
29#include "suggest/policyimpl/gesture/gesture_suggest_policy_factory.h"
30#include "suggest/policyimpl/typing/typing_suggest_policy_factory.h"
31#include "utils/int_array_view.h"
32#include "utils/log_utils.h"
33#include "utils/time_keeper.h"
34
35namespace latinime {
36
37const int Dictionary::HEADER_ATTRIBUTE_BUFFER_SIZE = 32;
38
39Dictionary::Dictionary(JNIEnv *env, DictionaryStructureWithBufferPolicy::StructurePolicyPtr
40        dictionaryStructureWithBufferPolicy)
41        : mDictionaryStructureWithBufferPolicy(std::move(dictionaryStructureWithBufferPolicy)),
42          mGestureSuggest(new Suggest(GestureSuggestPolicyFactory::getGestureSuggestPolicy())),
43          mTypingSuggest(new Suggest(TypingSuggestPolicyFactory::getTypingSuggestPolicy())) {
44    logDictionaryInfo(env);
45}
46
47void Dictionary::getSuggestions(ProximityInfo *proximityInfo, DicTraverseSession *traverseSession,
48        int *xcoordinates, int *ycoordinates, int *times, int *pointerIds, int *inputCodePoints,
49        int inputSize, const PrevWordsInfo *const prevWordsInfo,
50        const SuggestOptions *const suggestOptions, const float languageWeight,
51        SuggestionResults *const outSuggestionResults) const {
52    TimeKeeper::setCurrentTime();
53    traverseSession->init(this, prevWordsInfo, suggestOptions);
54    const auto &suggest = suggestOptions->isGesture() ? mGestureSuggest : mTypingSuggest;
55    suggest->getSuggestions(proximityInfo, traverseSession, xcoordinates,
56            ycoordinates, times, pointerIds, inputCodePoints, inputSize,
57            languageWeight, outSuggestionResults);
58    if (DEBUG_DICT) {
59        outSuggestionResults->dumpSuggestions();
60    }
61}
62
63Dictionary::NgramListenerForPrediction::NgramListenerForPrediction(
64        const PrevWordsInfo *const prevWordsInfo, const WordIdArrayView prevWordIds,
65        SuggestionResults *const suggestionResults,
66        const DictionaryStructureWithBufferPolicy *const dictStructurePolicy)
67    : mPrevWordsInfo(prevWordsInfo), mPrevWordIds(prevWordIds),
68      mSuggestionResults(suggestionResults), mDictStructurePolicy(dictStructurePolicy) {}
69
70void Dictionary::NgramListenerForPrediction::onVisitEntry(const int ngramProbability,
71        const int targetWordId) {
72    if (targetWordId == NOT_A_WORD_ID) {
73        return;
74    }
75    if (mPrevWordsInfo->isNthPrevWordBeginningOfSentence(1 /* n */)
76            && ngramProbability == NOT_A_PROBABILITY) {
77        return;
78    }
79    int targetWordCodePoints[MAX_WORD_LENGTH];
80    int unigramProbability = 0;
81    const int codePointCount = mDictStructurePolicy->
82            getCodePointsAndProbabilityAndReturnCodePointCount(targetWordId, MAX_WORD_LENGTH,
83                    targetWordCodePoints, &unigramProbability);
84    if (codePointCount <= 0) {
85        return;
86    }
87    const WordAttributes wordAttributes = mDictStructurePolicy->getWordAttributesInContext(
88            mPrevWordIds, targetWordId, nullptr /* multiBigramMap */);
89    mSuggestionResults->addPrediction(targetWordCodePoints, codePointCount,
90            wordAttributes.getProbability());
91}
92
93void Dictionary::getPredictions(const PrevWordsInfo *const prevWordsInfo,
94        SuggestionResults *const outSuggestionResults) const {
95    TimeKeeper::setCurrentTime();
96    WordIdArray<MAX_PREV_WORD_COUNT_FOR_N_GRAM> prevWordIdArray;
97    const WordIdArrayView prevWordIds = prevWordsInfo->getPrevWordIds(
98            mDictionaryStructureWithBufferPolicy.get(), &prevWordIdArray,
99            true /* tryLowerCaseSearch */);
100    NgramListenerForPrediction listener(prevWordsInfo, prevWordIds, outSuggestionResults,
101            mDictionaryStructureWithBufferPolicy.get());
102    mDictionaryStructureWithBufferPolicy->iterateNgramEntries(prevWordIds, &listener);
103}
104
105int Dictionary::getProbability(const CodePointArrayView codePoints) const {
106    return getNgramProbability(nullptr /* prevWordsInfo */, codePoints);
107}
108
109int Dictionary::getMaxProbabilityOfExactMatches(const CodePointArrayView codePoints) const {
110    TimeKeeper::setCurrentTime();
111    return DictionaryUtils::getMaxProbabilityOfExactMatches(
112            mDictionaryStructureWithBufferPolicy.get(), codePoints.data(), codePoints.size());
113}
114
115int Dictionary::getNgramProbability(const PrevWordsInfo *const prevWordsInfo,
116        const CodePointArrayView codePoints) const {
117    TimeKeeper::setCurrentTime();
118    const int wordId = mDictionaryStructureWithBufferPolicy->getWordId(codePoints,
119            false /* forceLowerCaseSearch */);
120    if (wordId == NOT_A_WORD_ID) return NOT_A_PROBABILITY;
121    if (!prevWordsInfo) {
122        return getDictionaryStructurePolicy()->getProbabilityOfWord(WordIdArrayView(), wordId);
123    }
124    WordIdArray<MAX_PREV_WORD_COUNT_FOR_N_GRAM> prevWordIdArray;
125    const WordIdArrayView prevWordIds = prevWordsInfo->getPrevWordIds
126            (mDictionaryStructureWithBufferPolicy.get(), &prevWordIdArray,
127            true /* tryLowerCaseSearch */);
128    return getDictionaryStructurePolicy()->getProbabilityOfWord(prevWordIds, wordId);
129}
130
131bool Dictionary::addUnigramEntry(const CodePointArrayView codePoints,
132        const UnigramProperty *const unigramProperty) {
133    if (unigramProperty->representsBeginningOfSentence()
134            && !mDictionaryStructureWithBufferPolicy->getHeaderStructurePolicy()
135                    ->supportsBeginningOfSentence()) {
136        AKLOGE("The dictionary doesn't support Beginning-of-Sentence.");
137        return false;
138    }
139    TimeKeeper::setCurrentTime();
140    return mDictionaryStructureWithBufferPolicy->addUnigramEntry(codePoints, unigramProperty);
141}
142
143bool Dictionary::removeUnigramEntry(const CodePointArrayView codePoints) {
144    TimeKeeper::setCurrentTime();
145    return mDictionaryStructureWithBufferPolicy->removeUnigramEntry(codePoints);
146}
147
148bool Dictionary::addNgramEntry(const PrevWordsInfo *const prevWordsInfo,
149        const BigramProperty *const bigramProperty) {
150    TimeKeeper::setCurrentTime();
151    return mDictionaryStructureWithBufferPolicy->addNgramEntry(prevWordsInfo, bigramProperty);
152}
153
154bool Dictionary::removeNgramEntry(const PrevWordsInfo *const prevWordsInfo,
155        const CodePointArrayView codePoints) {
156    TimeKeeper::setCurrentTime();
157    return mDictionaryStructureWithBufferPolicy->removeNgramEntry(prevWordsInfo, codePoints);
158}
159
160bool Dictionary::flush(const char *const filePath) {
161    TimeKeeper::setCurrentTime();
162    return mDictionaryStructureWithBufferPolicy->flush(filePath);
163}
164
165bool Dictionary::flushWithGC(const char *const filePath) {
166    TimeKeeper::setCurrentTime();
167    return mDictionaryStructureWithBufferPolicy->flushWithGC(filePath);
168}
169
170bool Dictionary::needsToRunGC(const bool mindsBlockByGC) {
171    TimeKeeper::setCurrentTime();
172    return mDictionaryStructureWithBufferPolicy->needsToRunGC(mindsBlockByGC);
173}
174
175void Dictionary::getProperty(const char *const query, const int queryLength, char *const outResult,
176        const int maxResultLength) {
177    TimeKeeper::setCurrentTime();
178    return mDictionaryStructureWithBufferPolicy->getProperty(query, queryLength, outResult,
179            maxResultLength);
180}
181
182const WordProperty Dictionary::getWordProperty(const CodePointArrayView codePoints) {
183    TimeKeeper::setCurrentTime();
184    return mDictionaryStructureWithBufferPolicy->getWordProperty(codePoints);
185}
186
187int Dictionary::getNextWordAndNextToken(const int token, int *const outCodePoints,
188        int *const outCodePointCount) {
189    TimeKeeper::setCurrentTime();
190    return mDictionaryStructureWithBufferPolicy->getNextWordAndNextToken(
191            token, outCodePoints, outCodePointCount);
192}
193
194void Dictionary::logDictionaryInfo(JNIEnv *const env) const {
195    int dictionaryIdCodePointBuffer[HEADER_ATTRIBUTE_BUFFER_SIZE];
196    int versionStringCodePointBuffer[HEADER_ATTRIBUTE_BUFFER_SIZE];
197    int dateStringCodePointBuffer[HEADER_ATTRIBUTE_BUFFER_SIZE];
198    const DictionaryHeaderStructurePolicy *const headerPolicy =
199            getDictionaryStructurePolicy()->getHeaderStructurePolicy();
200    headerPolicy->readHeaderValueOrQuestionMark("dictionary", dictionaryIdCodePointBuffer,
201            HEADER_ATTRIBUTE_BUFFER_SIZE);
202    headerPolicy->readHeaderValueOrQuestionMark("version", versionStringCodePointBuffer,
203            HEADER_ATTRIBUTE_BUFFER_SIZE);
204    headerPolicy->readHeaderValueOrQuestionMark("date", dateStringCodePointBuffer,
205            HEADER_ATTRIBUTE_BUFFER_SIZE);
206
207    char dictionaryIdCharBuffer[HEADER_ATTRIBUTE_BUFFER_SIZE];
208    char versionStringCharBuffer[HEADER_ATTRIBUTE_BUFFER_SIZE];
209    char dateStringCharBuffer[HEADER_ATTRIBUTE_BUFFER_SIZE];
210    intArrayToCharArray(dictionaryIdCodePointBuffer, HEADER_ATTRIBUTE_BUFFER_SIZE,
211            dictionaryIdCharBuffer, HEADER_ATTRIBUTE_BUFFER_SIZE);
212    intArrayToCharArray(versionStringCodePointBuffer, HEADER_ATTRIBUTE_BUFFER_SIZE,
213            versionStringCharBuffer, HEADER_ATTRIBUTE_BUFFER_SIZE);
214    intArrayToCharArray(dateStringCodePointBuffer, HEADER_ATTRIBUTE_BUFFER_SIZE,
215            dateStringCharBuffer, HEADER_ATTRIBUTE_BUFFER_SIZE);
216
217    LogUtils::logToJava(env,
218            "Dictionary info: dictionary = %s ; version = %s ; date = %s",
219            dictionaryIdCharBuffer, versionStringCharBuffer, dateStringCharBuffer);
220}
221
222} // namespace latinime
223