dictionary.cpp revision 2d57b3339ad5b4bbf0939858c36c7daf5e38a4cb
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/log_utils.h"
32#include "utils/time_keeper.h"
33
34namespace latinime {
35
36const int Dictionary::HEADER_ATTRIBUTE_BUFFER_SIZE = 32;
37
38Dictionary::Dictionary(JNIEnv *env, DictionaryStructureWithBufferPolicy::StructurePolicyPtr
39        dictionaryStructureWithBufferPolicy)
40        : mDictionaryStructureWithBufferPolicy(std::move(dictionaryStructureWithBufferPolicy)),
41          mGestureSuggest(new Suggest(GestureSuggestPolicyFactory::getGestureSuggestPolicy())),
42          mTypingSuggest(new Suggest(TypingSuggestPolicyFactory::getTypingSuggestPolicy())) {
43    logDictionaryInfo(env);
44}
45
46void Dictionary::getSuggestions(ProximityInfo *proximityInfo, DicTraverseSession *traverseSession,
47        int *xcoordinates, int *ycoordinates, int *times, int *pointerIds, int *inputCodePoints,
48        int inputSize, const PrevWordsInfo *const prevWordsInfo,
49        const SuggestOptions *const suggestOptions, const float languageWeight,
50        SuggestionResults *const outSuggestionResults) const {
51    TimeKeeper::setCurrentTime();
52    traverseSession->init(this, prevWordsInfo, suggestOptions);
53    const auto &suggest = suggestOptions->isGesture() ? mGestureSuggest : mTypingSuggest;
54    suggest->getSuggestions(proximityInfo, traverseSession, xcoordinates,
55            ycoordinates, times, pointerIds, inputCodePoints, inputSize,
56            languageWeight, outSuggestionResults);
57    if (DEBUG_DICT) {
58        outSuggestionResults->dumpSuggestions();
59    }
60}
61
62Dictionary::NgramListenerForPrediction::NgramListenerForPrediction(
63        const PrevWordsInfo *const prevWordsInfo, SuggestionResults *const suggestionResults,
64        const DictionaryStructureWithBufferPolicy *const dictStructurePolicy)
65    : mPrevWordsInfo(prevWordsInfo), mSuggestionResults(suggestionResults),
66      mDictStructurePolicy(dictStructurePolicy) {}
67
68void Dictionary::NgramListenerForPrediction::onVisitEntry(const int ngramProbability,
69        const int targetPtNodePos) {
70    if (targetPtNodePos == NOT_A_DICT_POS) {
71        return;
72    }
73    if (mPrevWordsInfo->isNthPrevWordBeginningOfSentence(1 /* n */)
74            && ngramProbability == NOT_A_PROBABILITY) {
75        return;
76    }
77    int targetWordCodePoints[MAX_WORD_LENGTH];
78    int unigramProbability = 0;
79    const int codePointCount = mDictStructurePolicy->
80            getCodePointsAndProbabilityAndReturnCodePointCount(targetPtNodePos,
81                    MAX_WORD_LENGTH, targetWordCodePoints, &unigramProbability);
82    if (codePointCount <= 0) {
83        return;
84    }
85    const int probability = mDictStructurePolicy->getProbability(
86            unigramProbability, ngramProbability);
87    mSuggestionResults->addPrediction(targetWordCodePoints, codePointCount, probability);
88}
89
90void Dictionary::getPredictions(const PrevWordsInfo *const prevWordsInfo,
91        SuggestionResults *const outSuggestionResults) const {
92    TimeKeeper::setCurrentTime();
93    NgramListenerForPrediction listener(prevWordsInfo, outSuggestionResults,
94            mDictionaryStructureWithBufferPolicy.get());
95    mDictionaryStructureWithBufferPolicy->iterateNgramEntries(prevWordsInfo, &listener);
96}
97
98int Dictionary::getProbability(const int *word, int length) const {
99    return getNgramProbability(nullptr /* prevWordsInfo */, word, length);
100}
101
102int Dictionary::getMaxProbabilityOfExactMatches(const int *word, int length) const {
103    TimeKeeper::setCurrentTime();
104    return DictionaryUtils::getMaxProbabilityOfExactMatches(
105            mDictionaryStructureWithBufferPolicy.get(), word, length);
106}
107
108int Dictionary::getNgramProbability(const PrevWordsInfo *const prevWordsInfo, const int *word,
109        int length) const {
110    TimeKeeper::setCurrentTime();
111    int nextWordPos = mDictionaryStructureWithBufferPolicy->getTerminalPtNodePositionOfWord(word,
112            length, false /* forceLowerCaseSearch */);
113    if (NOT_A_DICT_POS == nextWordPos) return NOT_A_PROBABILITY;
114    return getDictionaryStructurePolicy()->getProbabilityOfPtNode(prevWordsInfo, nextWordPos);
115}
116
117bool Dictionary::addUnigramEntry(const int *const word, const int length,
118        const UnigramProperty *const unigramProperty) {
119    if (unigramProperty->representsBeginningOfSentence()
120            && !mDictionaryStructureWithBufferPolicy->getHeaderStructurePolicy()
121                    ->supportsBeginningOfSentence()) {
122        AKLOGE("The dictionary doesn't support Beginning-of-Sentence.");
123        return false;
124    }
125    TimeKeeper::setCurrentTime();
126    return mDictionaryStructureWithBufferPolicy->addUnigramEntry(word, length, unigramProperty);
127}
128
129bool Dictionary::removeUnigramEntry(const int *const codePoints, const int codePointCount) {
130    TimeKeeper::setCurrentTime();
131    return mDictionaryStructureWithBufferPolicy->removeUnigramEntry(codePoints, codePointCount);
132}
133
134bool Dictionary::addNgramEntry(const PrevWordsInfo *const prevWordsInfo,
135        const BigramProperty *const bigramProperty) {
136    TimeKeeper::setCurrentTime();
137    return mDictionaryStructureWithBufferPolicy->addNgramEntry(prevWordsInfo, bigramProperty);
138}
139
140bool Dictionary::removeNgramEntry(const PrevWordsInfo *const prevWordsInfo,
141        const int *const word, const int length) {
142    TimeKeeper::setCurrentTime();
143    return mDictionaryStructureWithBufferPolicy->removeNgramEntry(prevWordsInfo, word, length);
144}
145
146bool Dictionary::flush(const char *const filePath) {
147    TimeKeeper::setCurrentTime();
148    return mDictionaryStructureWithBufferPolicy->flush(filePath);
149}
150
151bool Dictionary::flushWithGC(const char *const filePath) {
152    TimeKeeper::setCurrentTime();
153    return mDictionaryStructureWithBufferPolicy->flushWithGC(filePath);
154}
155
156bool Dictionary::needsToRunGC(const bool mindsBlockByGC) {
157    TimeKeeper::setCurrentTime();
158    return mDictionaryStructureWithBufferPolicy->needsToRunGC(mindsBlockByGC);
159}
160
161void Dictionary::getProperty(const char *const query, const int queryLength, char *const outResult,
162        const int maxResultLength) {
163    TimeKeeper::setCurrentTime();
164    return mDictionaryStructureWithBufferPolicy->getProperty(query, queryLength, outResult,
165            maxResultLength);
166}
167
168const WordProperty Dictionary::getWordProperty(const int *const codePoints,
169        const int codePointCount) {
170    TimeKeeper::setCurrentTime();
171    return mDictionaryStructureWithBufferPolicy->getWordProperty(
172            codePoints, codePointCount);
173}
174
175int Dictionary::getNextWordAndNextToken(const int token, int *const outCodePoints,
176        int *const outCodePointCount) {
177    TimeKeeper::setCurrentTime();
178    return mDictionaryStructureWithBufferPolicy->getNextWordAndNextToken(
179            token, outCodePoints, outCodePointCount);
180}
181
182void Dictionary::logDictionaryInfo(JNIEnv *const env) const {
183    int dictionaryIdCodePointBuffer[HEADER_ATTRIBUTE_BUFFER_SIZE];
184    int versionStringCodePointBuffer[HEADER_ATTRIBUTE_BUFFER_SIZE];
185    int dateStringCodePointBuffer[HEADER_ATTRIBUTE_BUFFER_SIZE];
186    const DictionaryHeaderStructurePolicy *const headerPolicy =
187            getDictionaryStructurePolicy()->getHeaderStructurePolicy();
188    headerPolicy->readHeaderValueOrQuestionMark("dictionary", dictionaryIdCodePointBuffer,
189            HEADER_ATTRIBUTE_BUFFER_SIZE);
190    headerPolicy->readHeaderValueOrQuestionMark("version", versionStringCodePointBuffer,
191            HEADER_ATTRIBUTE_BUFFER_SIZE);
192    headerPolicy->readHeaderValueOrQuestionMark("date", dateStringCodePointBuffer,
193            HEADER_ATTRIBUTE_BUFFER_SIZE);
194
195    char dictionaryIdCharBuffer[HEADER_ATTRIBUTE_BUFFER_SIZE];
196    char versionStringCharBuffer[HEADER_ATTRIBUTE_BUFFER_SIZE];
197    char dateStringCharBuffer[HEADER_ATTRIBUTE_BUFFER_SIZE];
198    intArrayToCharArray(dictionaryIdCodePointBuffer, HEADER_ATTRIBUTE_BUFFER_SIZE,
199            dictionaryIdCharBuffer, HEADER_ATTRIBUTE_BUFFER_SIZE);
200    intArrayToCharArray(versionStringCodePointBuffer, HEADER_ATTRIBUTE_BUFFER_SIZE,
201            versionStringCharBuffer, HEADER_ATTRIBUTE_BUFFER_SIZE);
202    intArrayToCharArray(dateStringCodePointBuffer, HEADER_ATTRIBUTE_BUFFER_SIZE,
203            dateStringCharBuffer, HEADER_ATTRIBUTE_BUFFER_SIZE);
204
205    LogUtils::logToJava(env,
206            "Dictionary info: dictionary = %s ; version = %s ; date = %s",
207            dictionaryIdCharBuffer, versionStringCharBuffer, dateStringCharBuffer);
208}
209
210} // namespace latinime
211