diff options
Diffstat (limited to 'native/jni/src')
8 files changed, 27 insertions, 7 deletions
diff --git a/native/jni/src/suggest/policyimpl/dictionary/dynamic_patricia_trie_gc_event_listeners.cpp b/native/jni/src/suggest/policyimpl/dictionary/dynamic_patricia_trie_gc_event_listeners.cpp index a17a0acf6..5724c5d88 100644 --- a/native/jni/src/suggest/policyimpl/dictionary/dynamic_patricia_trie_gc_event_listeners.cpp +++ b/native/jni/src/suggest/policyimpl/dictionary/dynamic_patricia_trie_gc_event_listeners.cpp @@ -39,7 +39,7 @@ bool DynamicPatriciaTrieGcEventListeners return false; } if (!ForgettingCurveUtils::isValidEncodedProbability(newProbability)) { - isUselessPtNode = false; + isUselessPtNode = true; } } if (mChildrenValue > 0) { diff --git a/native/jni/src/suggest/policyimpl/dictionary/dynamic_patricia_trie_gc_event_listeners.h b/native/jni/src/suggest/policyimpl/dictionary/dynamic_patricia_trie_gc_event_listeners.h index 3ca2f2a01..9755120b0 100644 --- a/native/jni/src/suggest/policyimpl/dictionary/dynamic_patricia_trie_gc_event_listeners.h +++ b/native/jni/src/suggest/policyimpl/dictionary/dynamic_patricia_trie_gc_event_listeners.h @@ -60,6 +60,7 @@ class DynamicPatriciaTrieGcEventListeners { bool onDescend(const int ptNodeArrayPos) { mValueStack.push_back(0); + mChildrenValue = 0; return true; } diff --git a/native/jni/src/suggest/policyimpl/dictionary/dynamic_patricia_trie_policy.cpp b/native/jni/src/suggest/policyimpl/dictionary/dynamic_patricia_trie_policy.cpp index 31e3fb42f..3d07c9d6c 100644 --- a/native/jni/src/suggest/policyimpl/dictionary/dynamic_patricia_trie_policy.cpp +++ b/native/jni/src/suggest/policyimpl/dictionary/dynamic_patricia_trie_policy.cpp @@ -37,6 +37,8 @@ namespace latinime { // BinaryDictionaryDecayingTests. const char *const DynamicPatriciaTriePolicy::UNIGRAM_COUNT_QUERY = "UNIGRAM_COUNT"; const char *const DynamicPatriciaTriePolicy::BIGRAM_COUNT_QUERY = "BIGRAM_COUNT"; +const char *const DynamicPatriciaTriePolicy::MAX_UNIGRAM_COUNT_QUERY = "MAX_UNIGRAM_COUNT"; +const char *const DynamicPatriciaTriePolicy::MAX_BIGRAM_COUNT_QUERY = "MAX_BIGRAM_COUNT"; const char *const DynamicPatriciaTriePolicy::SET_NEEDS_TO_DECAY_FOR_TESTING_QUERY = "SET_NEEDS_TO_DECAY_FOR_TESTING"; const int DynamicPatriciaTriePolicy::MAX_DICT_EXTENDED_REGION_SIZE = 1024 * 1024; @@ -355,6 +357,14 @@ void DynamicPatriciaTriePolicy::getProperty(const char *const query, char *const snprintf(outResult, maxResultLength, "%d", mUnigramCount); } else if (strncmp(query, BIGRAM_COUNT_QUERY, maxResultLength) == 0) { snprintf(outResult, maxResultLength, "%d", mBigramCount); + } else if (strncmp(query, MAX_UNIGRAM_COUNT_QUERY, maxResultLength) == 0) { + snprintf(outResult, maxResultLength, "%d", + mHeaderPolicy.isDecayingDict() ? ForgettingCurveUtils::MAX_UNIGRAM_COUNT : + DynamicPatriciaTrieWritingHelper::MAX_DICTIONARY_SIZE); + } else if (strncmp(query, MAX_BIGRAM_COUNT_QUERY, maxResultLength) == 0) { + snprintf(outResult, maxResultLength, "%d", + mHeaderPolicy.isDecayingDict() ? ForgettingCurveUtils::MAX_BIGRAM_COUNT : + DynamicPatriciaTrieWritingHelper::MAX_DICTIONARY_SIZE); } else if (strncmp(query, SET_NEEDS_TO_DECAY_FOR_TESTING_QUERY, maxResultLength) == 0) { mNeedsToDecayForTesting = true; } diff --git a/native/jni/src/suggest/policyimpl/dictionary/dynamic_patricia_trie_policy.h b/native/jni/src/suggest/policyimpl/dictionary/dynamic_patricia_trie_policy.h index 903f65e8e..be97ee1a5 100644 --- a/native/jni/src/suggest/policyimpl/dictionary/dynamic_patricia_trie_policy.h +++ b/native/jni/src/suggest/policyimpl/dictionary/dynamic_patricia_trie_policy.h @@ -102,6 +102,8 @@ class DynamicPatriciaTriePolicy : public DictionaryStructureWithBufferPolicy { static const char *const UNIGRAM_COUNT_QUERY; static const char *const BIGRAM_COUNT_QUERY; + static const char *const MAX_UNIGRAM_COUNT_QUERY; + static const char *const MAX_BIGRAM_COUNT_QUERY; static const char *const SET_NEEDS_TO_DECAY_FOR_TESTING_QUERY; static const int MAX_DICT_EXTENDED_REGION_SIZE; static const int MIN_DICT_SIZE_TO_REFUSE_DYNAMIC_OPERATIONS; diff --git a/native/jni/src/suggest/policyimpl/dictionary/dynamic_patricia_trie_reading_helper.cpp b/native/jni/src/suggest/policyimpl/dictionary/dynamic_patricia_trie_reading_helper.cpp index 601ee663b..f108c219f 100644 --- a/native/jni/src/suggest/policyimpl/dictionary/dynamic_patricia_trie_reading_helper.cpp +++ b/native/jni/src/suggest/policyimpl/dictionary/dynamic_patricia_trie_reading_helper.cpp @@ -93,6 +93,12 @@ bool DynamicPatriciaTrieReadingHelper::traverseAllPtNodesInPtNodeArrayLevelPreor if (!listener->onDescend(getPosOfLastPtNodeArrayHead())) { return false; } + if (isEnd()) { + // Empty dictionary. Needs to notify the listener of the tail of empty PtNode array. + if (!listener->onReadingPtNodeArrayTail()) { + return false; + } + } pushReadingStateToStack(); while (!isEnd()) { if (alreadyVisitedAllPtNodesInArray) { diff --git a/native/jni/src/suggest/policyimpl/dictionary/dynamic_patricia_trie_reading_helper.h b/native/jni/src/suggest/policyimpl/dictionary/dynamic_patricia_trie_reading_helper.h index 512a4d818..a71c06971 100644 --- a/native/jni/src/suggest/policyimpl/dictionary/dynamic_patricia_trie_reading_helper.h +++ b/native/jni/src/suggest/policyimpl/dictionary/dynamic_patricia_trie_reading_helper.h @@ -279,7 +279,9 @@ class DynamicPatriciaTrieReadingHelper { } else { mReadingState = mReadingStateStack.back(); mReadingStateStack.pop_back(); - fetchPtNodeInfo(); + if (!isEnd()) { + fetchPtNodeInfo(); + } } } }; diff --git a/native/jni/src/suggest/policyimpl/dictionary/utils/forgetting_curve_utils.cpp b/native/jni/src/suggest/policyimpl/dictionary/utils/forgetting_curve_utils.cpp index 19ca35481..1632fd072 100644 --- a/native/jni/src/suggest/policyimpl/dictionary/utils/forgetting_curve_utils.cpp +++ b/native/jni/src/suggest/policyimpl/dictionary/utils/forgetting_curve_utils.cpp @@ -93,8 +93,7 @@ void ForgettingCurveUtils::TimeKeeper::setCurrentTime() { for (int i = 0; i < decayIterationCount; ++i) { const float currentRate = static_cast<float>(currentEncodedProbability) / static_cast<float>(MAX_ENCODED_PROBABILITY); - const float thresholdToDecay = MIN_PROBABILITY_TO_DECAY - + (1.0f - MIN_PROBABILITY_TO_DECAY) * currentRate; + const float thresholdToDecay = (1.0f - MIN_PROBABILITY_TO_DECAY) * currentRate; const float randValue = static_cast<float>(rand()) / static_cast<float>(RAND_MAX); if (thresholdToDecay < randValue) { currentEncodedProbability = max(currentEncodedProbability - ENCODED_PROBABILITY_STEP, diff --git a/native/jni/src/suggest/policyimpl/typing/scoring_params.cpp b/native/jni/src/suggest/policyimpl/typing/scoring_params.cpp index 66637ac4b..104eb2a7a 100644 --- a/native/jni/src/suggest/policyimpl/typing/scoring_params.cpp +++ b/native/jni/src/suggest/policyimpl/typing/scoring_params.cpp @@ -35,7 +35,7 @@ const float ScoringParams::OMISSION_COST = 0.4676f; const float ScoringParams::OMISSION_COST_SAME_CHAR = 0.399f; const float ScoringParams::OMISSION_COST_FIRST_CHAR = 0.5256f; const float ScoringParams::INSERTION_COST = 0.7248f; -const float ScoringParams::TERMINAL_INSERTION_COST = 0.9828f; +const float ScoringParams::TERMINAL_INSERTION_COST = 0.8128f; const float ScoringParams::INSERTION_COST_SAME_CHAR = 0.5508f; const float ScoringParams::INSERTION_COST_PROXIMITY_CHAR = 0.674f; const float ScoringParams::INSERTION_COST_FIRST_CHAR = 0.639f; @@ -43,10 +43,10 @@ const float ScoringParams::TRANSPOSITION_COST = 0.5608f; const float ScoringParams::SPACE_SUBSTITUTION_COST = 0.339f; const float ScoringParams::ADDITIONAL_PROXIMITY_COST = 0.4576f; const float ScoringParams::SUBSTITUTION_COST = 0.3806f; -const float ScoringParams::COST_NEW_WORD = 0.0292f; +const float ScoringParams::COST_NEW_WORD = 0.0312f; const float ScoringParams::COST_SECOND_OR_LATER_WORD_FIRST_CHAR_UPPERCASE = 0.3224f; const float ScoringParams::DISTANCE_WEIGHT_LANGUAGE = 1.1214f; -const float ScoringParams::COST_FIRST_LOOKAHEAD = 0.4786f; +const float ScoringParams::COST_FIRST_LOOKAHEAD = 0.4836f; const float ScoringParams::COST_LOOKAHEAD = 0.00624f; const float ScoringParams::HAS_PROXIMITY_TERMINAL_COST = 0.06836f; const float ScoringParams::HAS_EDIT_CORRECTION_TERMINAL_COST = 0.0362f; |