dictionary.cpp revision 1229879e7c5892e818ab53b3c2162a158cc5e177
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 62void Dictionary::getPredictions(const PrevWordsInfo *const prevWordsInfo, 63 SuggestionResults *const outSuggestionResults) const { 64 TimeKeeper::setCurrentTime(); 65 int unigramProbability = 0; 66 int bigramCodePoints[MAX_WORD_LENGTH]; 67 BinaryDictionaryBigramsIterator bigramsIt = prevWordsInfo->getBigramsIteratorForPrediction( 68 mDictionaryStructureWithBufferPolicy.get()); 69 while (bigramsIt.hasNext()) { 70 bigramsIt.next(); 71 if (bigramsIt.getBigramPos() == NOT_A_DICT_POS) { 72 continue; 73 } 74 if (prevWordsInfo->isNthPrevWordBeginningOfSentence(1 /* n */) 75 && bigramsIt.getProbability() == NOT_A_PROBABILITY) { 76 continue; 77 } 78 const int codePointCount = mDictionaryStructureWithBufferPolicy-> 79 getCodePointsAndProbabilityAndReturnCodePointCount(bigramsIt.getBigramPos(), 80 MAX_WORD_LENGTH, bigramCodePoints, &unigramProbability); 81 if (codePointCount <= 0) { 82 continue; 83 } 84 const int probability = mDictionaryStructureWithBufferPolicy->getProbability( 85 unigramProbability, bigramsIt.getProbability()); 86 outSuggestionResults->addPrediction(bigramCodePoints, codePointCount, probability); 87 } 88} 89 90int Dictionary::getProbability(const int *word, int length) const { 91 return getNgramProbability(nullptr /* prevWordsInfo */, word, length); 92} 93 94int Dictionary::getMaxProbabilityOfExactMatches(const int *word, int length) const { 95 TimeKeeper::setCurrentTime(); 96 return DictionaryUtils::getMaxProbabilityOfExactMatches( 97 mDictionaryStructureWithBufferPolicy.get(), word, length); 98} 99 100int Dictionary::getNgramProbability(const PrevWordsInfo *const prevWordsInfo, const int *word, 101 int length) const { 102 TimeKeeper::setCurrentTime(); 103 int nextWordPos = mDictionaryStructureWithBufferPolicy->getTerminalPtNodePositionOfWord(word, 104 length, false /* forceLowerCaseSearch */); 105 if (NOT_A_DICT_POS == nextWordPos) return NOT_A_PROBABILITY; 106 return getDictionaryStructurePolicy()->getProbabilityOfPtNode(prevWordsInfo, nextWordPos); 107} 108 109bool Dictionary::addUnigramEntry(const int *const word, const int length, 110 const UnigramProperty *const unigramProperty) { 111 if (unigramProperty->representsBeginningOfSentence() 112 && !mDictionaryStructureWithBufferPolicy->getHeaderStructurePolicy() 113 ->supportsBeginningOfSentence()) { 114 AKLOGE("The dictionary doesn't support Beginning-of-Sentence."); 115 return false; 116 } 117 TimeKeeper::setCurrentTime(); 118 return mDictionaryStructureWithBufferPolicy->addUnigramEntry(word, length, unigramProperty); 119} 120 121bool Dictionary::removeUnigramEntry(const int *const codePoints, const int codePointCount) { 122 TimeKeeper::setCurrentTime(); 123 return mDictionaryStructureWithBufferPolicy->removeUnigramEntry(codePoints, codePointCount); 124} 125 126bool Dictionary::addNgramEntry(const PrevWordsInfo *const prevWordsInfo, 127 const BigramProperty *const bigramProperty) { 128 TimeKeeper::setCurrentTime(); 129 return mDictionaryStructureWithBufferPolicy->addNgramEntry(prevWordsInfo, bigramProperty); 130} 131 132bool Dictionary::removeNgramEntry(const PrevWordsInfo *const prevWordsInfo, 133 const int *const word, const int length) { 134 TimeKeeper::setCurrentTime(); 135 return mDictionaryStructureWithBufferPolicy->removeNgramEntry(prevWordsInfo, word, length); 136} 137 138bool Dictionary::flush(const char *const filePath) { 139 TimeKeeper::setCurrentTime(); 140 return mDictionaryStructureWithBufferPolicy->flush(filePath); 141} 142 143bool Dictionary::flushWithGC(const char *const filePath) { 144 TimeKeeper::setCurrentTime(); 145 return mDictionaryStructureWithBufferPolicy->flushWithGC(filePath); 146} 147 148bool Dictionary::needsToRunGC(const bool mindsBlockByGC) { 149 TimeKeeper::setCurrentTime(); 150 return mDictionaryStructureWithBufferPolicy->needsToRunGC(mindsBlockByGC); 151} 152 153void Dictionary::getProperty(const char *const query, const int queryLength, char *const outResult, 154 const int maxResultLength) { 155 TimeKeeper::setCurrentTime(); 156 return mDictionaryStructureWithBufferPolicy->getProperty(query, queryLength, outResult, 157 maxResultLength); 158} 159 160const WordProperty Dictionary::getWordProperty(const int *const codePoints, 161 const int codePointCount) { 162 TimeKeeper::setCurrentTime(); 163 return mDictionaryStructureWithBufferPolicy->getWordProperty( 164 codePoints, codePointCount); 165} 166 167int Dictionary::getNextWordAndNextToken(const int token, int *const outCodePoints, 168 int *const outCodePointCount) { 169 TimeKeeper::setCurrentTime(); 170 return mDictionaryStructureWithBufferPolicy->getNextWordAndNextToken( 171 token, outCodePoints, outCodePointCount); 172} 173 174void Dictionary::logDictionaryInfo(JNIEnv *const env) const { 175 int dictionaryIdCodePointBuffer[HEADER_ATTRIBUTE_BUFFER_SIZE]; 176 int versionStringCodePointBuffer[HEADER_ATTRIBUTE_BUFFER_SIZE]; 177 int dateStringCodePointBuffer[HEADER_ATTRIBUTE_BUFFER_SIZE]; 178 const DictionaryHeaderStructurePolicy *const headerPolicy = 179 getDictionaryStructurePolicy()->getHeaderStructurePolicy(); 180 headerPolicy->readHeaderValueOrQuestionMark("dictionary", dictionaryIdCodePointBuffer, 181 HEADER_ATTRIBUTE_BUFFER_SIZE); 182 headerPolicy->readHeaderValueOrQuestionMark("version", versionStringCodePointBuffer, 183 HEADER_ATTRIBUTE_BUFFER_SIZE); 184 headerPolicy->readHeaderValueOrQuestionMark("date", dateStringCodePointBuffer, 185 HEADER_ATTRIBUTE_BUFFER_SIZE); 186 187 char dictionaryIdCharBuffer[HEADER_ATTRIBUTE_BUFFER_SIZE]; 188 char versionStringCharBuffer[HEADER_ATTRIBUTE_BUFFER_SIZE]; 189 char dateStringCharBuffer[HEADER_ATTRIBUTE_BUFFER_SIZE]; 190 intArrayToCharArray(dictionaryIdCodePointBuffer, HEADER_ATTRIBUTE_BUFFER_SIZE, 191 dictionaryIdCharBuffer, HEADER_ATTRIBUTE_BUFFER_SIZE); 192 intArrayToCharArray(versionStringCodePointBuffer, HEADER_ATTRIBUTE_BUFFER_SIZE, 193 versionStringCharBuffer, HEADER_ATTRIBUTE_BUFFER_SIZE); 194 intArrayToCharArray(dateStringCodePointBuffer, HEADER_ATTRIBUTE_BUFFER_SIZE, 195 dateStringCharBuffer, HEADER_ATTRIBUTE_BUFFER_SIZE); 196 197 LogUtils::logToJava(env, 198 "Dictionary info: dictionary = %s ; version = %s ; date = %s", 199 dictionaryIdCharBuffer, versionStringCharBuffer, dateStringCharBuffer); 200} 201 202} // namespace latinime 203