diff options
Diffstat (limited to 'native/jni/src')
18 files changed, 124 insertions, 120 deletions
diff --git a/native/jni/src/suggest/core/dictionary/dictionary.cpp b/native/jni/src/suggest/core/dictionary/dictionary.cpp index 7a69d3ceb..697e99ffb 100644 --- a/native/jni/src/suggest/core/dictionary/dictionary.cpp +++ b/native/jni/src/suggest/core/dictionary/dictionary.cpp @@ -23,7 +23,7 @@ #include "suggest/core/policy/dictionary_header_structure_policy.h" #include "suggest/core/result/suggestion_results.h" #include "suggest/core/session/dic_traverse_session.h" -#include "suggest/core/session/prev_words_info.h" +#include "suggest/core/session/ngram_context.h" #include "suggest/core/suggest.h" #include "suggest/core/suggest_options.h" #include "suggest/policyimpl/gesture/gesture_suggest_policy_factory.h" @@ -46,11 +46,11 @@ Dictionary::Dictionary(JNIEnv *env, DictionaryStructureWithBufferPolicy::Structu void Dictionary::getSuggestions(ProximityInfo *proximityInfo, DicTraverseSession *traverseSession, int *xcoordinates, int *ycoordinates, int *times, int *pointerIds, int *inputCodePoints, - int inputSize, const PrevWordsInfo *const prevWordsInfo, + int inputSize, const NgramContext *const ngramContext, const SuggestOptions *const suggestOptions, const float weightOfLangModelVsSpatialModel, SuggestionResults *const outSuggestionResults) const { TimeKeeper::setCurrentTime(); - traverseSession->init(this, prevWordsInfo, suggestOptions); + traverseSession->init(this, ngramContext, suggestOptions); const auto &suggest = suggestOptions->isGesture() ? mGestureSuggest : mTypingSuggest; suggest->getSuggestions(proximityInfo, traverseSession, xcoordinates, ycoordinates, times, pointerIds, inputCodePoints, inputSize, @@ -58,10 +58,10 @@ void Dictionary::getSuggestions(ProximityInfo *proximityInfo, DicTraverseSession } Dictionary::NgramListenerForPrediction::NgramListenerForPrediction( - const PrevWordsInfo *const prevWordsInfo, const WordIdArrayView prevWordIds, + const NgramContext *const ngramContext, const WordIdArrayView prevWordIds, SuggestionResults *const suggestionResults, const DictionaryStructureWithBufferPolicy *const dictStructurePolicy) - : mPrevWordsInfo(prevWordsInfo), mPrevWordIds(prevWordIds), + : mNgramContext(ngramContext), mPrevWordIds(prevWordIds), mSuggestionResults(suggestionResults), mDictStructurePolicy(dictStructurePolicy) {} void Dictionary::NgramListenerForPrediction::onVisitEntry(const int ngramProbability, @@ -69,7 +69,7 @@ void Dictionary::NgramListenerForPrediction::onVisitEntry(const int ngramProbabi if (targetWordId == NOT_A_WORD_ID) { return; } - if (mPrevWordsInfo->isNthPrevWordBeginningOfSentence(1 /* n */) + if (mNgramContext->isNthPrevWordBeginningOfSentence(1 /* n */) && ngramProbability == NOT_A_PROBABILITY) { return; } @@ -85,20 +85,20 @@ void Dictionary::NgramListenerForPrediction::onVisitEntry(const int ngramProbabi wordAttributes.getProbability()); } -void Dictionary::getPredictions(const PrevWordsInfo *const prevWordsInfo, +void Dictionary::getPredictions(const NgramContext *const ngramContext, SuggestionResults *const outSuggestionResults) const { TimeKeeper::setCurrentTime(); WordIdArray<MAX_PREV_WORD_COUNT_FOR_N_GRAM> prevWordIdArray; - const WordIdArrayView prevWordIds = prevWordsInfo->getPrevWordIds( + const WordIdArrayView prevWordIds = ngramContext->getPrevWordIds( mDictionaryStructureWithBufferPolicy.get(), &prevWordIdArray, true /* tryLowerCaseSearch */); - NgramListenerForPrediction listener(prevWordsInfo, prevWordIds, outSuggestionResults, + NgramListenerForPrediction listener(ngramContext, prevWordIds, outSuggestionResults, mDictionaryStructureWithBufferPolicy.get()); mDictionaryStructureWithBufferPolicy->iterateNgramEntries(prevWordIds, &listener); } int Dictionary::getProbability(const CodePointArrayView codePoints) const { - return getNgramProbability(nullptr /* prevWordsInfo */, codePoints); + return getNgramProbability(nullptr /* ngramContext */, codePoints); } int Dictionary::getMaxProbabilityOfExactMatches(const CodePointArrayView codePoints) const { @@ -107,18 +107,18 @@ int Dictionary::getMaxProbabilityOfExactMatches(const CodePointArrayView codePoi mDictionaryStructureWithBufferPolicy.get(), codePoints); } -int Dictionary::getNgramProbability(const PrevWordsInfo *const prevWordsInfo, +int Dictionary::getNgramProbability(const NgramContext *const ngramContext, const CodePointArrayView codePoints) const { TimeKeeper::setCurrentTime(); const int wordId = mDictionaryStructureWithBufferPolicy->getWordId(codePoints, false /* forceLowerCaseSearch */); if (wordId == NOT_A_WORD_ID) return NOT_A_PROBABILITY; - if (!prevWordsInfo) { + if (!ngramContext) { return getDictionaryStructurePolicy()->getProbabilityOfWord(WordIdArrayView(), wordId); } WordIdArray<MAX_PREV_WORD_COUNT_FOR_N_GRAM> prevWordIdArray; - const WordIdArrayView prevWordIds = prevWordsInfo->getPrevWordIds - (mDictionaryStructureWithBufferPolicy.get(), &prevWordIdArray, + const WordIdArrayView prevWordIds = ngramContext->getPrevWordIds( + mDictionaryStructureWithBufferPolicy.get(), &prevWordIdArray, true /* tryLowerCaseSearch */); return getDictionaryStructurePolicy()->getProbabilityOfWord(prevWordIds, wordId); } @@ -140,24 +140,24 @@ bool Dictionary::removeUnigramEntry(const CodePointArrayView codePoints) { return mDictionaryStructureWithBufferPolicy->removeUnigramEntry(codePoints); } -bool Dictionary::addNgramEntry(const PrevWordsInfo *const prevWordsInfo, +bool Dictionary::addNgramEntry(const NgramContext *const ngramContext, const NgramProperty *const ngramProperty) { TimeKeeper::setCurrentTime(); - return mDictionaryStructureWithBufferPolicy->addNgramEntry(prevWordsInfo, ngramProperty); + return mDictionaryStructureWithBufferPolicy->addNgramEntry(ngramContext, ngramProperty); } -bool Dictionary::removeNgramEntry(const PrevWordsInfo *const prevWordsInfo, +bool Dictionary::removeNgramEntry(const NgramContext *const ngramContext, const CodePointArrayView codePoints) { TimeKeeper::setCurrentTime(); - return mDictionaryStructureWithBufferPolicy->removeNgramEntry(prevWordsInfo, codePoints); + return mDictionaryStructureWithBufferPolicy->removeNgramEntry(ngramContext, codePoints); } -bool Dictionary::updateCounter(const PrevWordsInfo *const prevWordsInfo, +bool Dictionary::updateEntriesForWordWithNgramContext(const NgramContext *const ngramContext, const CodePointArrayView codePoints, const bool isValidWord, const HistoricalInfo historicalInfo) { TimeKeeper::setCurrentTime(); - return mDictionaryStructureWithBufferPolicy->updateCounter(prevWordsInfo, codePoints, - isValidWord, historicalInfo); + return mDictionaryStructureWithBufferPolicy->updateEntriesForWordWithNgramContext(ngramContext, + codePoints, isValidWord, historicalInfo); } bool Dictionary::flush(const char *const filePath) { diff --git a/native/jni/src/suggest/core/dictionary/dictionary.h b/native/jni/src/suggest/core/dictionary/dictionary.h index a58dbfbd7..843aec473 100644 --- a/native/jni/src/suggest/core/dictionary/dictionary.h +++ b/native/jni/src/suggest/core/dictionary/dictionary.h @@ -33,7 +33,7 @@ namespace latinime { class DictionaryStructureWithBufferPolicy; class DicTraverseSession; -class PrevWordsInfo; +class NgramContext; class ProximityInfo; class SuggestionResults; class SuggestOptions; @@ -66,18 +66,18 @@ class Dictionary { void getSuggestions(ProximityInfo *proximityInfo, DicTraverseSession *traverseSession, int *xcoordinates, int *ycoordinates, int *times, int *pointerIds, int *inputCodePoints, - int inputSize, const PrevWordsInfo *const prevWordsInfo, + int inputSize, const NgramContext *const ngramContext, const SuggestOptions *const suggestOptions, const float weightOfLangModelVsSpatialModel, SuggestionResults *const outSuggestionResults) const; - void getPredictions(const PrevWordsInfo *const prevWordsInfo, + void getPredictions(const NgramContext *const ngramContext, SuggestionResults *const outSuggestionResults) const; int getProbability(const CodePointArrayView codePoints) const; int getMaxProbabilityOfExactMatches(const CodePointArrayView codePoints) const; - int getNgramProbability(const PrevWordsInfo *const prevWordsInfo, + int getNgramProbability(const NgramContext *const ngramContext, const CodePointArrayView codePoints) const; bool addUnigramEntry(const CodePointArrayView codePoints, @@ -85,13 +85,13 @@ class Dictionary { bool removeUnigramEntry(const CodePointArrayView codePoints); - bool addNgramEntry(const PrevWordsInfo *const prevWordsInfo, + bool addNgramEntry(const NgramContext *const ngramContext, const NgramProperty *const ngramProperty); - bool removeNgramEntry(const PrevWordsInfo *const prevWordsInfo, + bool removeNgramEntry(const NgramContext *const ngramContext, const CodePointArrayView codePoints); - bool updateCounter(const PrevWordsInfo *const prevWordsInfo, + bool updateEntriesForWordWithNgramContext(const NgramContext *const ngramContext, const CodePointArrayView codePoints, const bool isValidWord, const HistoricalInfo historicalInfo); @@ -123,7 +123,7 @@ class Dictionary { class NgramListenerForPrediction : public NgramListener { public: - NgramListenerForPrediction(const PrevWordsInfo *const prevWordsInfo, + NgramListenerForPrediction(const NgramContext *const ngramContext, const WordIdArrayView prevWordIds, SuggestionResults *const suggestionResults, const DictionaryStructureWithBufferPolicy *const dictStructurePolicy); virtual void onVisitEntry(const int ngramProbability, const int targetWordId); @@ -131,7 +131,7 @@ class Dictionary { private: DISALLOW_IMPLICIT_CONSTRUCTORS(NgramListenerForPrediction); - const PrevWordsInfo *const mPrevWordsInfo; + const NgramContext *const mNgramContext; const WordIdArrayView mPrevWordIds; SuggestionResults *const mSuggestionResults; const DictionaryStructureWithBufferPolicy *const mDictStructurePolicy; diff --git a/native/jni/src/suggest/core/dictionary/dictionary_utils.cpp b/native/jni/src/suggest/core/dictionary/dictionary_utils.cpp index b85f3622a..9573c37bc 100644 --- a/native/jni/src/suggest/core/dictionary/dictionary_utils.cpp +++ b/native/jni/src/suggest/core/dictionary/dictionary_utils.cpp @@ -21,7 +21,7 @@ #include "suggest/core/dicnode/dic_node_vector.h" #include "suggest/core/dictionary/dictionary.h" #include "suggest/core/dictionary/digraph_utils.h" -#include "suggest/core/session/prev_words_info.h" +#include "suggest/core/session/ngram_context.h" #include "suggest/core/policy/dictionary_structure_with_buffer_policy.h" #include "utils/int_array_view.h" @@ -33,10 +33,10 @@ namespace latinime { std::vector<DicNode> current; std::vector<DicNode> next; - // No prev words information. - PrevWordsInfo emptyPrevWordsInfo; + // No ngram context. + NgramContext emptyNgramContext; WordIdArray<MAX_PREV_WORD_COUNT_FOR_N_GRAM> prevWordIdArray; - const WordIdArrayView prevWordIds = emptyPrevWordsInfo.getPrevWordIds( + const WordIdArrayView prevWordIds = emptyNgramContext.getPrevWordIds( dictionaryStructurePolicy, &prevWordIdArray, false /* tryLowerCaseSearch */); current.emplace_back(); DicNodeUtils::initAsRoot(dictionaryStructurePolicy, prevWordIds, ¤t.front()); diff --git a/native/jni/src/suggest/core/policy/dictionary_structure_with_buffer_policy.h b/native/jni/src/suggest/core/policy/dictionary_structure_with_buffer_policy.h index 6624b7921..ceda5c03f 100644 --- a/native/jni/src/suggest/core/policy/dictionary_structure_with_buffer_policy.h +++ b/native/jni/src/suggest/core/policy/dictionary_structure_with_buffer_policy.h @@ -33,7 +33,7 @@ class DicNodeVector; class DictionaryHeaderStructurePolicy; class MultiBigramMap; class NgramListener; -class PrevWordsInfo; +class NgramContext; class UnigramProperty; /* @@ -81,15 +81,15 @@ class DictionaryStructureWithBufferPolicy { virtual bool removeUnigramEntry(const CodePointArrayView wordCodePoints) = 0; // Returns whether the update was success or not. - virtual bool addNgramEntry(const PrevWordsInfo *const prevWordsInfo, + virtual bool addNgramEntry(const NgramContext *const ngramContext, const NgramProperty *const ngramProperty) = 0; // Returns whether the update was success or not. - virtual bool removeNgramEntry(const PrevWordsInfo *const prevWordsInfo, + virtual bool removeNgramEntry(const NgramContext *const ngramContext, const CodePointArrayView wordCodePoints) = 0; // Returns whether the update was success or not. - virtual bool updateCounter(const PrevWordsInfo *const prevWordsInfo, + virtual bool updateEntriesForWordWithNgramContext(const NgramContext *const ngramContext, const CodePointArrayView wordCodePoints, const bool isValidWord, const HistoricalInfo historicalInfo) = 0; diff --git a/native/jni/src/suggest/core/session/dic_traverse_session.cpp b/native/jni/src/suggest/core/session/dic_traverse_session.cpp index b4d01d0f0..52dc2f86c 100644 --- a/native/jni/src/suggest/core/session/dic_traverse_session.cpp +++ b/native/jni/src/suggest/core/session/dic_traverse_session.cpp @@ -20,7 +20,7 @@ #include "suggest/core/dictionary/dictionary.h" #include "suggest/core/policy/dictionary_header_structure_policy.h" #include "suggest/core/policy/dictionary_structure_with_buffer_policy.h" -#include "suggest/core/session/prev_words_info.h" +#include "suggest/core/session/ngram_context.h" namespace latinime { @@ -30,12 +30,12 @@ const int DicTraverseSession::DICTIONARY_SIZE_THRESHOLD_TO_USE_LARGE_CACHE_FOR_S 256 * 1024; void DicTraverseSession::init(const Dictionary *const dictionary, - const PrevWordsInfo *const prevWordsInfo, const SuggestOptions *const suggestOptions) { + const NgramContext *const ngramContext, const SuggestOptions *const suggestOptions) { mDictionary = dictionary; mMultiWordCostMultiplier = getDictionaryStructurePolicy()->getHeaderStructurePolicy() ->getMultiWordCostMultiplier(); mSuggestOptions = suggestOptions; - mPrevWordIdCount = prevWordsInfo->getPrevWordIds(getDictionaryStructurePolicy(), + mPrevWordIdCount = ngramContext->getPrevWordIds(getDictionaryStructurePolicy(), &mPrevWordIdArray, true /* tryLowerCaseSearch */).size(); } diff --git a/native/jni/src/suggest/core/session/dic_traverse_session.h b/native/jni/src/suggest/core/session/dic_traverse_session.h index 9f841aa3c..bc53167f0 100644 --- a/native/jni/src/suggest/core/session/dic_traverse_session.h +++ b/native/jni/src/suggest/core/session/dic_traverse_session.h @@ -30,7 +30,7 @@ namespace latinime { class Dictionary; class DictionaryStructureWithBufferPolicy; -class PrevWordsInfo; +class NgramContext; class ProximityInfo; class SuggestOptions; @@ -61,7 +61,7 @@ class DicTraverseSession { // Non virtual inline destructor -- never inherit this class AK_FORCE_INLINE ~DicTraverseSession() {} - void init(const Dictionary *dictionary, const PrevWordsInfo *const prevWordsInfo, + void init(const Dictionary *dictionary, const NgramContext *const ngramContext, const SuggestOptions *const suggestOptions); // TODO: Remove and merge into init void setupForGetSuggestions(const ProximityInfo *pInfo, const int *inputCodePoints, diff --git a/native/jni/src/suggest/core/session/prev_words_info.h b/native/jni/src/suggest/core/session/ngram_context.h index 553d5ad07..64c71410f 100644 --- a/native/jni/src/suggest/core/session/prev_words_info.h +++ b/native/jni/src/suggest/core/session/ngram_context.h @@ -14,8 +14,8 @@ * limitations under the License. */ -#ifndef LATINIME_PREV_WORDS_INFO_H -#define LATINIME_PREV_WORDS_INFO_H +#ifndef LATINIME_NGRAM_CONTEXT_H +#define LATINIME_NGRAM_CONTEXT_H #include <array> @@ -26,25 +26,26 @@ namespace latinime { -class PrevWordsInfo { +// Rename to NgramContext. +class NgramContext { public: // No prev word information. - PrevWordsInfo() : mPrevWordCount(0) { + NgramContext() : mPrevWordCount(0) { clear(); } - PrevWordsInfo(const PrevWordsInfo &prevWordsInfo) - : mPrevWordCount(prevWordsInfo.mPrevWordCount) { + NgramContext(const NgramContext &ngramContext) + : mPrevWordCount(ngramContext.mPrevWordCount) { for (size_t i = 0; i < mPrevWordCount; ++i) { - mPrevWordCodePointCount[i] = prevWordsInfo.mPrevWordCodePointCount[i]; - memmove(mPrevWordCodePoints[i], prevWordsInfo.mPrevWordCodePoints[i], + mPrevWordCodePointCount[i] = ngramContext.mPrevWordCodePointCount[i]; + memmove(mPrevWordCodePoints[i], ngramContext.mPrevWordCodePoints[i], sizeof(mPrevWordCodePoints[i][0]) * mPrevWordCodePointCount[i]); - mIsBeginningOfSentence[i] = prevWordsInfo.mIsBeginningOfSentence[i]; + mIsBeginningOfSentence[i] = ngramContext.mIsBeginningOfSentence[i]; } } // Construct from previous words. - PrevWordsInfo(const int prevWordCodePoints[][MAX_WORD_LENGTH], + NgramContext(const int prevWordCodePoints[][MAX_WORD_LENGTH], const int *const prevWordCodePointCount, const bool *const isBeginningOfSentence, const size_t prevWordCount) : mPrevWordCount(std::min(NELEMS(mPrevWordCodePoints), prevWordCount)) { @@ -61,7 +62,7 @@ class PrevWordsInfo { } // Construct from a previous word. - PrevWordsInfo(const int *const prevWordCodePoints, const int prevWordCodePointCount, + NgramContext(const int *const prevWordCodePoints, const int prevWordCodePointCount, const bool isBeginningOfSentence) : mPrevWordCount(1) { clear(); if (prevWordCodePointCount > MAX_WORD_LENGTH || !prevWordCodePoints) { @@ -78,8 +79,8 @@ class PrevWordsInfo { } // TODO: Remove. - const PrevWordsInfo getTrimmedPrevWordsInfo(const size_t maxPrevWordCount) const { - return PrevWordsInfo(mPrevWordCodePoints, mPrevWordCodePointCount, mIsBeginningOfSentence, + const NgramContext getTrimmedNgramContext(const size_t maxPrevWordCount) const { + return NgramContext(mPrevWordCodePoints, mPrevWordCodePointCount, mIsBeginningOfSentence, std::min(mPrevWordCount, maxPrevWordCount)); } @@ -122,7 +123,7 @@ class PrevWordsInfo { } private: - DISALLOW_ASSIGNMENT_OPERATOR(PrevWordsInfo); + DISALLOW_ASSIGNMENT_OPERATOR(NgramContext); static int getWordId(const DictionaryStructureWithBufferPolicy *const dictStructurePolicy, const int *const wordCodePoints, const int wordCodePointCount, @@ -165,4 +166,4 @@ class PrevWordsInfo { bool mIsBeginningOfSentence[MAX_PREV_WORD_COUNT_FOR_N_GRAM]; }; } // namespace latinime -#endif // LATINIME_PREV_WORDS_INFO_H +#endif // LATINIME_NGRAM_CONTEXT_H diff --git a/native/jni/src/suggest/policyimpl/dictionary/structure/backward/v402/ver4_patricia_trie_node_writer.cpp b/native/jni/src/suggest/policyimpl/dictionary/structure/backward/v402/ver4_patricia_trie_node_writer.cpp index 8d169743c..6243f14cc 100644 --- a/native/jni/src/suggest/policyimpl/dictionary/structure/backward/v402/ver4_patricia_trie_node_writer.cpp +++ b/native/jni/src/suggest/policyimpl/dictionary/structure/backward/v402/ver4_patricia_trie_node_writer.cpp @@ -310,7 +310,7 @@ bool Ver4PatriciaTrieNodeWriter::addShortcutTarget(const PtNodeParams *const ptN const int shortcutProbability) { if (!mShortcutPolicy->addNewShortcut(ptNodeParams->getTerminalId(), targetCodePoints, targetCodePointCount, shortcutProbability)) { - AKLOGE("Cannot add new shortuct entry. terminalId: %d", ptNodeParams->getTerminalId()); + AKLOGE("Cannot add new shortcut entry. terminalId: %d", ptNodeParams->getTerminalId()); return false; } if (!ptNodeParams->hasShortcutTargets()) { diff --git a/native/jni/src/suggest/policyimpl/dictionary/structure/backward/v402/ver4_patricia_trie_policy.cpp b/native/jni/src/suggest/policyimpl/dictionary/structure/backward/v402/ver4_patricia_trie_policy.cpp index 36eafa1e9..1c61bd4b9 100644 --- a/native/jni/src/suggest/policyimpl/dictionary/structure/backward/v402/ver4_patricia_trie_policy.cpp +++ b/native/jni/src/suggest/policyimpl/dictionary/structure/backward/v402/ver4_patricia_trie_policy.cpp @@ -33,7 +33,7 @@ #include "suggest/core/dictionary/property/ngram_property.h" #include "suggest/core/dictionary/property/unigram_property.h" #include "suggest/core/dictionary/property/word_property.h" -#include "suggest/core/session/prev_words_info.h" +#include "suggest/core/session/ngram_context.h" #include "suggest/policyimpl/dictionary/structure/pt_common/dynamic_pt_reading_helper.h" #include "suggest/policyimpl/dictionary/structure/backward/v402/ver4_patricia_trie_node_reader.h" #include "suggest/policyimpl/dictionary/utils/forgetting_curve_utils.h" @@ -338,7 +338,7 @@ bool Ver4PatriciaTriePolicy::removeUnigramEntry(const CodePointArrayView wordCod return mNodeWriter.suppressUnigramEntry(&ptNodeParams); } -bool Ver4PatriciaTriePolicy::addNgramEntry(const PrevWordsInfo *const prevWordsInfo, +bool Ver4PatriciaTriePolicy::addNgramEntry(const NgramContext *const ngramContext, const NgramProperty *const ngramProperty) { if (!mBuffers->isUpdatable()) { AKLOGI("Warning: addNgramEntry() is called for non-updatable dictionary."); @@ -349,8 +349,8 @@ bool Ver4PatriciaTriePolicy::addNgramEntry(const PrevWordsInfo *const prevWordsI mDictBuffer->getTailPosition()); return false; } - if (!prevWordsInfo->isValid()) { - AKLOGE("prev words info is not valid for adding n-gram entry to the dictionary."); + if (!ngramContext->isValid()) { + AKLOGE("Ngram context is not valid for adding n-gram entry to the dictionary."); return false; } if (ngramProperty->getTargetCodePoints()->size() > MAX_WORD_LENGTH) { @@ -359,23 +359,23 @@ bool Ver4PatriciaTriePolicy::addNgramEntry(const PrevWordsInfo *const prevWordsI return false; } WordIdArray<MAX_PREV_WORD_COUNT_FOR_N_GRAM> prevWordIdArray; - const WordIdArrayView prevWordIds = prevWordsInfo->getPrevWordIds(this, &prevWordIdArray, + const WordIdArrayView prevWordIds = ngramContext->getPrevWordIds(this, &prevWordIdArray, false /* tryLowerCaseSearch */); if (prevWordIds.empty()) { return false; } if (prevWordIds[0] == NOT_A_WORD_ID) { - if (prevWordsInfo->isNthPrevWordBeginningOfSentence(1 /* n */)) { + if (ngramContext->isNthPrevWordBeginningOfSentence(1 /* n */)) { const UnigramProperty beginningOfSentenceUnigramProperty( true /* representsBeginningOfSentence */, true /* isNotAWord */, false /* isBlacklisted */, MAX_PROBABILITY /* probability */, HistoricalInfo()); - if (!addUnigramEntry(prevWordsInfo->getNthPrevWordCodePoints(1 /* n */), + if (!addUnigramEntry(ngramContext->getNthPrevWordCodePoints(1 /* n */), &beginningOfSentenceUnigramProperty)) { AKLOGE("Cannot add unigram entry for the beginning-of-sentence."); return false; } // Refresh word ids. - prevWordsInfo->getPrevWordIds(this, &prevWordIdArray, false /* tryLowerCaseSearch */); + ngramContext->getPrevWordIds(this, &prevWordIdArray, false /* tryLowerCaseSearch */); } else { return false; } @@ -399,7 +399,7 @@ bool Ver4PatriciaTriePolicy::addNgramEntry(const PrevWordsInfo *const prevWordsI } } -bool Ver4PatriciaTriePolicy::removeNgramEntry(const PrevWordsInfo *const prevWordsInfo, +bool Ver4PatriciaTriePolicy::removeNgramEntry(const NgramContext *const ngramContext, const CodePointArrayView wordCodePoints) { if (!mBuffers->isUpdatable()) { AKLOGI("Warning: removeNgramEntry() is called for non-updatable dictionary."); @@ -410,8 +410,8 @@ bool Ver4PatriciaTriePolicy::removeNgramEntry(const PrevWordsInfo *const prevWor mDictBuffer->getTailPosition()); return false; } - if (!prevWordsInfo->isValid()) { - AKLOGE("prev words info is not valid for removing n-gram entry form the dictionary."); + if (!ngramContext->isValid()) { + AKLOGE("Ngram context is not valid for removing n-gram entry form the dictionary."); return false; } if (wordCodePoints.size() > MAX_WORD_LENGTH) { @@ -419,7 +419,7 @@ bool Ver4PatriciaTriePolicy::removeNgramEntry(const PrevWordsInfo *const prevWor wordCodePoints.size()); } WordIdArray<MAX_PREV_WORD_COUNT_FOR_N_GRAM> prevWordIdArray; - const WordIdArrayView prevWordIds = prevWordsInfo->getPrevWordIds(this, &prevWordIdArray, + const WordIdArrayView prevWordIds = ngramContext->getPrevWordIds(this, &prevWordIdArray, false /* tryLowerCaseSerch */); if (prevWordIds.firstOrDefault(NOT_A_WORD_ID) == NOT_A_WORD_ID) { return false; @@ -440,26 +440,27 @@ bool Ver4PatriciaTriePolicy::removeNgramEntry(const PrevWordsInfo *const prevWor } -bool Ver4PatriciaTriePolicy::updateCounter(const PrevWordsInfo *const prevWordsInfo, - const CodePointArrayView wordCodePoints, const bool isValidWord, - const HistoricalInfo historicalInfo) { +bool Ver4PatriciaTriePolicy::updateEntriesForWordWithNgramContext( + const NgramContext *const ngramContext, const CodePointArrayView wordCodePoints, + const bool isValidWord, const HistoricalInfo historicalInfo) { if (!mBuffers->isUpdatable()) { - AKLOGI("Warning: updateCounter() is called for non-updatable dictionary."); + AKLOGI("Warning: updateEntriesForWordWithNgramContext() is called for non-updatable " + "dictionary."); return false; } const int probability = isValidWord ? DUMMY_PROBABILITY_FOR_VALID_WORDS : NOT_A_PROBABILITY; const UnigramProperty unigramProperty(false /* representsBeginningOfSentence */, false /* isNotAWord */, false /*isBlacklisted*/, probability, historicalInfo); if (!addUnigramEntry(wordCodePoints, &unigramProperty)) { - AKLOGE("Cannot update unigarm entry in updateCounter()."); + AKLOGE("Cannot update unigarm entry in updateEntriesForWordWithNgramContext()."); return false; } - const int probabilityForNgram = prevWordsInfo->isNthPrevWordBeginningOfSentence(1 /* n */) + const int probabilityForNgram = ngramContext->isNthPrevWordBeginningOfSentence(1 /* n */) ? NOT_A_PROBABILITY : probability; const NgramProperty ngramProperty(wordCodePoints.toVector(), probabilityForNgram, historicalInfo); - if (!addNgramEntry(prevWordsInfo, &ngramProperty)) { - AKLOGE("Cannot update unigarm entry in updateCounter()."); + if (!addNgramEntry(ngramContext, &ngramProperty)) { + AKLOGE("Cannot update unigarm entry in updateEntriesForWordWithNgramContext()."); return false; } return true; diff --git a/native/jni/src/suggest/policyimpl/dictionary/structure/backward/v402/ver4_patricia_trie_policy.h b/native/jni/src/suggest/policyimpl/dictionary/structure/backward/v402/ver4_patricia_trie_policy.h index b82563e61..4aa399c3e 100644 --- a/native/jni/src/suggest/policyimpl/dictionary/structure/backward/v402/ver4_patricia_trie_policy.h +++ b/native/jni/src/suggest/policyimpl/dictionary/structure/backward/v402/ver4_patricia_trie_policy.h @@ -112,13 +112,13 @@ class Ver4PatriciaTriePolicy : public DictionaryStructureWithBufferPolicy { bool removeUnigramEntry(const CodePointArrayView wordCodePoints); - bool addNgramEntry(const PrevWordsInfo *const prevWordsInfo, + bool addNgramEntry(const NgramContext *const ngramContext, const NgramProperty *const ngramProperty); - bool removeNgramEntry(const PrevWordsInfo *const prevWordsInfo, + bool removeNgramEntry(const NgramContext *const ngramContext, const CodePointArrayView wordCodePoints); - bool updateCounter(const PrevWordsInfo *const prevWordsInfo, + bool updateEntriesForWordWithNgramContext(const NgramContext *const ngramContext, const CodePointArrayView wordCodePoints, const bool isValidWord, const HistoricalInfo historicalInfo); diff --git a/native/jni/src/suggest/policyimpl/dictionary/structure/v2/patricia_trie_policy.cpp b/native/jni/src/suggest/policyimpl/dictionary/structure/v2/patricia_trie_policy.cpp index d3d684bfa..b7f1199c5 100644 --- a/native/jni/src/suggest/policyimpl/dictionary/structure/v2/patricia_trie_policy.cpp +++ b/native/jni/src/suggest/policyimpl/dictionary/structure/v2/patricia_trie_policy.cpp @@ -23,7 +23,7 @@ #include "suggest/core/dictionary/binary_dictionary_bigrams_iterator.h" #include "suggest/core/dictionary/multi_bigram_map.h" #include "suggest/core/dictionary/ngram_listener.h" -#include "suggest/core/session/prev_words_info.h" +#include "suggest/core/session/ngram_context.h" #include "suggest/policyimpl/dictionary/structure/pt_common/dynamic_pt_reading_helper.h" #include "suggest/policyimpl/dictionary/structure/pt_common/patricia_trie_reading_utils.h" #include "suggest/policyimpl/dictionary/utils/probability_utils.h" diff --git a/native/jni/src/suggest/policyimpl/dictionary/structure/v2/patricia_trie_policy.h b/native/jni/src/suggest/policyimpl/dictionary/structure/v2/patricia_trie_policy.h index 32a95bb6c..b17681388 100644 --- a/native/jni/src/suggest/policyimpl/dictionary/structure/v2/patricia_trie_policy.h +++ b/native/jni/src/suggest/policyimpl/dictionary/structure/v2/patricia_trie_policy.h @@ -93,25 +93,26 @@ class PatriciaTriePolicy : public DictionaryStructureWithBufferPolicy { return false; } - bool addNgramEntry(const PrevWordsInfo *const prevWordsInfo, + bool addNgramEntry(const NgramContext *const ngramContext, const NgramProperty *const ngramProperty) { // This method should not be called for non-updatable dictionary. AKLOGI("Warning: addNgramEntry() is called for non-updatable dictionary."); return false; } - bool removeNgramEntry(const PrevWordsInfo *const prevWordsInfo, + bool removeNgramEntry(const NgramContext *const ngramContext, const CodePointArrayView wordCodePoints) { // This method should not be called for non-updatable dictionary. AKLOGI("Warning: removeNgramEntry() is called for non-updatable dictionary."); return false; } - bool updateCounter(const PrevWordsInfo *const prevWordsInfo, + bool updateEntriesForWordWithNgramContext(const NgramContext *const ngramContext, const CodePointArrayView wordCodePoints, const bool isValidWord, const HistoricalInfo historicalInfo) { // This method should not be called for non-updatable dictionary. - AKLOGI("Warning: updateCounter() is called for non-updatable dictionary."); + AKLOGI("Warning: updateEntriesForWordWithNgramContext() is called for non-updatable " + "dictionary."); return false; } diff --git a/native/jni/src/suggest/policyimpl/dictionary/structure/v2/ver2_patricia_trie_node_reader.cpp b/native/jni/src/suggest/policyimpl/dictionary/structure/v2/ver2_patricia_trie_node_reader.cpp index dc0ed96d0..90d4687dd 100644 --- a/native/jni/src/suggest/policyimpl/dictionary/structure/v2/ver2_patricia_trie_node_reader.cpp +++ b/native/jni/src/suggest/policyimpl/dictionary/structure/v2/ver2_patricia_trie_node_reader.cpp @@ -37,7 +37,7 @@ const PtNodeParams Ver2ParticiaTrieNodeReader::fetchPtNodeParamsInBufferFromPtNo int shortcutPos = NOT_A_DICT_POS; int bigramPos = NOT_A_DICT_POS; int siblingPos = NOT_A_DICT_POS; - PatriciaTrieReadingUtils::readPtNodeInfo(mBuffer.data(), ptNodePos, mShortuctPolicy, + PatriciaTrieReadingUtils::readPtNodeInfo(mBuffer.data(), ptNodePos, mShortcutPolicy, mBigramPolicy, mCodePointTable, &flags, &mergedNodeCodePointCount, mergedNodeCodePoints, &probability, &childrenPos, &shortcutPos, &bigramPos, &siblingPos); if (mergedNodeCodePointCount <= 0) { diff --git a/native/jni/src/suggest/policyimpl/dictionary/structure/v2/ver2_patricia_trie_node_reader.h b/native/jni/src/suggest/policyimpl/dictionary/structure/v2/ver2_patricia_trie_node_reader.h index 24ec5bcca..838d37314 100644 --- a/native/jni/src/suggest/policyimpl/dictionary/structure/v2/ver2_patricia_trie_node_reader.h +++ b/native/jni/src/suggest/policyimpl/dictionary/structure/v2/ver2_patricia_trie_node_reader.h @@ -35,7 +35,7 @@ class Ver2ParticiaTrieNodeReader : public PtNodeReader { const DictionaryBigramsStructurePolicy *const bigramPolicy, const DictionaryShortcutsStructurePolicy *const shortcutPolicy, const int *const codePointTable) - : mBuffer(buffer), mBigramPolicy(bigramPolicy), mShortuctPolicy(shortcutPolicy), + : mBuffer(buffer), mBigramPolicy(bigramPolicy), mShortcutPolicy(shortcutPolicy), mCodePointTable(codePointTable) {} virtual const PtNodeParams fetchPtNodeParamsInBufferFromPtNodePos(const int ptNodePos) const; @@ -45,7 +45,7 @@ class Ver2ParticiaTrieNodeReader : public PtNodeReader { const ReadOnlyByteArrayView mBuffer; const DictionaryBigramsStructurePolicy *const mBigramPolicy; - const DictionaryShortcutsStructurePolicy *const mShortuctPolicy; + const DictionaryShortcutsStructurePolicy *const mShortcutPolicy; const int *const mCodePointTable; }; } // namespace latinime diff --git a/native/jni/src/suggest/policyimpl/dictionary/structure/v4/ver4_patricia_trie_node_writer.cpp b/native/jni/src/suggest/policyimpl/dictionary/structure/v4/ver4_patricia_trie_node_writer.cpp index f13512d5a..d28006ae9 100644 --- a/native/jni/src/suggest/policyimpl/dictionary/structure/v4/ver4_patricia_trie_node_writer.cpp +++ b/native/jni/src/suggest/policyimpl/dictionary/structure/v4/ver4_patricia_trie_node_writer.cpp @@ -285,7 +285,7 @@ bool Ver4PatriciaTrieNodeWriter::addShortcutTarget(const PtNodeParams *const ptN const int shortcutProbability) { if (!mShortcutPolicy->addNewShortcut(ptNodeParams->getTerminalId(), targetCodePoints, targetCodePointCount, shortcutProbability)) { - AKLOGE("Cannot add new shortuct entry. terminalId: %d", ptNodeParams->getTerminalId()); + AKLOGE("Cannot add new shortcut entry. terminalId: %d", ptNodeParams->getTerminalId()); return false; } return true; diff --git a/native/jni/src/suggest/policyimpl/dictionary/structure/v4/ver4_patricia_trie_policy.cpp b/native/jni/src/suggest/policyimpl/dictionary/structure/v4/ver4_patricia_trie_policy.cpp index 036197c41..445bbe07e 100644 --- a/native/jni/src/suggest/policyimpl/dictionary/structure/v4/ver4_patricia_trie_policy.cpp +++ b/native/jni/src/suggest/policyimpl/dictionary/structure/v4/ver4_patricia_trie_policy.cpp @@ -26,7 +26,7 @@ #include "suggest/core/dictionary/property/ngram_property.h" #include "suggest/core/dictionary/property/unigram_property.h" #include "suggest/core/dictionary/property/word_property.h" -#include "suggest/core/session/prev_words_info.h" +#include "suggest/core/session/ngram_context.h" #include "suggest/policyimpl/dictionary/structure/pt_common/dynamic_pt_reading_helper.h" #include "suggest/policyimpl/dictionary/structure/v4/ver4_patricia_trie_node_reader.h" #include "suggest/policyimpl/dictionary/utils/forgetting_curve_utils.h" @@ -266,7 +266,7 @@ bool Ver4PatriciaTriePolicy::removeUnigramEntry(const CodePointArrayView wordCod return true; } -bool Ver4PatriciaTriePolicy::addNgramEntry(const PrevWordsInfo *const prevWordsInfo, +bool Ver4PatriciaTriePolicy::addNgramEntry(const NgramContext *const ngramContext, const NgramProperty *const ngramProperty) { if (!mBuffers->isUpdatable()) { AKLOGI("Warning: addNgramEntry() is called for non-updatable dictionary."); @@ -277,8 +277,8 @@ bool Ver4PatriciaTriePolicy::addNgramEntry(const PrevWordsInfo *const prevWordsI mDictBuffer->getTailPosition()); return false; } - if (!prevWordsInfo->isValid()) { - AKLOGE("prev words info is not valid for adding n-gram entry to the dictionary."); + if (!ngramContext->isValid()) { + AKLOGE("Ngram context is not valid for adding n-gram entry to the dictionary."); return false; } if (ngramProperty->getTargetCodePoints()->size() > MAX_WORD_LENGTH) { @@ -287,7 +287,7 @@ bool Ver4PatriciaTriePolicy::addNgramEntry(const PrevWordsInfo *const prevWordsI return false; } WordIdArray<MAX_PREV_WORD_COUNT_FOR_N_GRAM> prevWordIdArray; - const WordIdArrayView prevWordIds = prevWordsInfo->getPrevWordIds(this, &prevWordIdArray, + const WordIdArrayView prevWordIds = ngramContext->getPrevWordIds(this, &prevWordIdArray, false /* tryLowerCaseSearch */); if (prevWordIds.empty()) { return false; @@ -296,19 +296,19 @@ bool Ver4PatriciaTriePolicy::addNgramEntry(const PrevWordsInfo *const prevWordsI if (prevWordIds[i] != NOT_A_WORD_ID) { continue; } - if (!prevWordsInfo->isNthPrevWordBeginningOfSentence(i + 1 /* n */)) { + if (!ngramContext->isNthPrevWordBeginningOfSentence(i + 1 /* n */)) { return false; } const UnigramProperty beginningOfSentenceUnigramProperty( true /* representsBeginningOfSentence */, true /* isNotAWord */, false /* isBlacklisted */, MAX_PROBABILITY /* probability */, HistoricalInfo()); - if (!addUnigramEntry(prevWordsInfo->getNthPrevWordCodePoints(1 /* n */), + if (!addUnigramEntry(ngramContext->getNthPrevWordCodePoints(1 /* n */), &beginningOfSentenceUnigramProperty)) { AKLOGE("Cannot add unigram entry for the beginning-of-sentence."); return false; } // Refresh word ids. - prevWordsInfo->getPrevWordIds(this, &prevWordIdArray, false /* tryLowerCaseSearch */); + ngramContext->getPrevWordIds(this, &prevWordIdArray, false /* tryLowerCaseSearch */); } const int wordId = getWordId(CodePointArrayView(*ngramProperty->getTargetCodePoints()), false /* forceLowerCaseSearch */); @@ -326,7 +326,7 @@ bool Ver4PatriciaTriePolicy::addNgramEntry(const PrevWordsInfo *const prevWordsI } } -bool Ver4PatriciaTriePolicy::removeNgramEntry(const PrevWordsInfo *const prevWordsInfo, +bool Ver4PatriciaTriePolicy::removeNgramEntry(const NgramContext *const ngramContext, const CodePointArrayView wordCodePoints) { if (!mBuffers->isUpdatable()) { AKLOGI("Warning: removeNgramEntry() is called for non-updatable dictionary."); @@ -337,8 +337,8 @@ bool Ver4PatriciaTriePolicy::removeNgramEntry(const PrevWordsInfo *const prevWor mDictBuffer->getTailPosition()); return false; } - if (!prevWordsInfo->isValid()) { - AKLOGE("prev words info is not valid for removing n-gram entry form the dictionary."); + if (!ngramContext->isValid()) { + AKLOGE("Ngram context is not valid for removing n-gram entry form the dictionary."); return false; } if (wordCodePoints.size() > MAX_WORD_LENGTH) { @@ -346,7 +346,7 @@ bool Ver4PatriciaTriePolicy::removeNgramEntry(const PrevWordsInfo *const prevWor wordCodePoints.size()); } WordIdArray<MAX_PREV_WORD_COUNT_FOR_N_GRAM> prevWordIdArray; - const WordIdArrayView prevWordIds = prevWordsInfo->getPrevWordIds(this, &prevWordIdArray, + const WordIdArrayView prevWordIds = ngramContext->getPrevWordIds(this, &prevWordIdArray, false /* tryLowerCaseSerch */); if (prevWordIds.empty() || prevWordIds.contains(NOT_A_WORD_ID)) { return false; @@ -363,29 +363,30 @@ bool Ver4PatriciaTriePolicy::removeNgramEntry(const PrevWordsInfo *const prevWor } } -bool Ver4PatriciaTriePolicy::updateCounter(const PrevWordsInfo *const prevWordsInfo, - const CodePointArrayView wordCodePoints, const bool isValidWord, - const HistoricalInfo historicalInfo) { +bool Ver4PatriciaTriePolicy::updateEntriesForWordWithNgramContext( + const NgramContext *const ngramContext, const CodePointArrayView wordCodePoints, + const bool isValidWord, const HistoricalInfo historicalInfo) { if (!mBuffers->isUpdatable()) { - AKLOGI("Warning: updateCounter() is called for non-updatable dictionary."); + AKLOGI("Warning: updateEntriesForWordWithNgramContext() is called for non-updatable " + "dictionary."); return false; } // TODO: Have count up method in language model dict content. const int probability = isValidWord ? DUMMY_PROBABILITY_FOR_VALID_WORDS : NOT_A_PROBABILITY; const UnigramProperty unigramProperty(false /* representsBeginningOfSentence */, - false /* isNotAWord */, false /*isBlacklisted*/, probability, historicalInfo); + false /* isNotAWord */, false /* isBlacklisted */, probability, historicalInfo); if (!addUnigramEntry(wordCodePoints, &unigramProperty)) { - AKLOGE("Cannot update unigarm entry in updateCounter()."); + AKLOGE("Cannot update unigarm entry in updateEntriesForWordWithNgramContext()."); return false; } - const int probabilityForNgram = prevWordsInfo->isNthPrevWordBeginningOfSentence(1 /* n */) + const int probabilityForNgram = ngramContext->isNthPrevWordBeginningOfSentence(1 /* n */) ? NOT_A_PROBABILITY : probability; const NgramProperty ngramProperty(wordCodePoints.toVector(), probabilityForNgram, historicalInfo); - for (size_t i = 1; i <= prevWordsInfo->getPrevWordCount(); ++i) { - const PrevWordsInfo trimmedPrevWordsInfo(prevWordsInfo->getTrimmedPrevWordsInfo(i)); - if (!addNgramEntry(&trimmedPrevWordsInfo, &ngramProperty)) { - AKLOGE("Cannot update ngram entry in updateCounter()."); + for (size_t i = 1; i <= ngramContext->getPrevWordCount(); ++i) { + const NgramContext trimmedNgramContext(ngramContext->getTrimmedNgramContext(i)); + if (!addNgramEntry(&trimmedNgramContext, &ngramProperty)) { + AKLOGE("Cannot update ngram entry in updateEntriesForWordWithNgramContext()."); return false; } } diff --git a/native/jni/src/suggest/policyimpl/dictionary/structure/v4/ver4_patricia_trie_policy.h b/native/jni/src/suggest/policyimpl/dictionary/structure/v4/ver4_patricia_trie_policy.h index 662bb8d4b..60e30f209 100644 --- a/native/jni/src/suggest/policyimpl/dictionary/structure/v4/ver4_patricia_trie_policy.h +++ b/native/jni/src/suggest/policyimpl/dictionary/structure/v4/ver4_patricia_trie_policy.h @@ -92,13 +92,13 @@ class Ver4PatriciaTriePolicy : public DictionaryStructureWithBufferPolicy { bool removeUnigramEntry(const CodePointArrayView wordCodePoints); - bool addNgramEntry(const PrevWordsInfo *const prevWordsInfo, + bool addNgramEntry(const NgramContext *const ngramContext, const NgramProperty *const ngramProperty); - bool removeNgramEntry(const PrevWordsInfo *const prevWordsInfo, + bool removeNgramEntry(const NgramContext *const ngramContext, const CodePointArrayView wordCodePoints); - bool updateCounter(const PrevWordsInfo *const prevWordsInfo, + bool updateEntriesForWordWithNgramContext(const NgramContext *const ngramContext, const CodePointArrayView wordCodePoints, const bool isValidWord, const HistoricalInfo historicalInfo); diff --git a/native/jni/src/utils/jni_data_utils.h b/native/jni/src/utils/jni_data_utils.h index 235a03bba..25cc41742 100644 --- a/native/jni/src/utils/jni_data_utils.h +++ b/native/jni/src/utils/jni_data_utils.h @@ -21,7 +21,7 @@ #include "defines.h" #include "jni.h" -#include "suggest/core/session/prev_words_info.h" +#include "suggest/core/session/ngram_context.h" #include "suggest/core/policy/dictionary_header_structure_policy.h" #include "suggest/policyimpl/dictionary/header/header_read_write_utils.h" #include "utils/char_utils.h" @@ -96,7 +96,7 @@ class JniDataUtils { } } - static PrevWordsInfo constructPrevWordsInfo(JNIEnv *env, jobjectArray prevWordCodePointArrays, + static NgramContext constructNgramContext(JNIEnv *env, jobjectArray prevWordCodePointArrays, jbooleanArray isBeginningOfSentenceArray, const size_t prevWordCount) { int prevWordCodePoints[MAX_PREV_WORD_COUNT_FOR_N_GRAM][MAX_WORD_LENGTH]; int prevWordCodePointCount[MAX_PREV_WORD_COUNT_FOR_N_GRAM]; @@ -119,7 +119,7 @@ class JniDataUtils { &isBeginningOfSentenceBoolean); isBeginningOfSentence[i] = isBeginningOfSentenceBoolean == JNI_TRUE; } - return PrevWordsInfo(prevWordCodePoints, prevWordCodePointCount, isBeginningOfSentence, + return NgramContext(prevWordCodePoints, prevWordCodePointCount, isBeginningOfSentence, prevWordCount); } |