mirror of
https://gitlab.futo.org/keyboard/latinime.git
synced 2024-09-28 14:54:30 +01:00
Remove BigramDictionary form native code.
Bug: 14425059 Change-Id: I48d1a4ae6a4317afaaf67748060618f8f9485d40
This commit is contained in:
parent
d58a07666e
commit
8681bef03c
@ -26,7 +26,6 @@ LATIN_IME_CORE_SRC_FILES := \
|
||||
dic_node_utils.cpp \
|
||||
dic_nodes_cache.cpp) \
|
||||
$(addprefix suggest/core/dictionary/, \
|
||||
bigram_dictionary.cpp \
|
||||
dictionary.cpp \
|
||||
dictionary_utils.cpp \
|
||||
digraph_utils.cpp \
|
||||
|
@ -1,112 +0,0 @@
|
||||
/*
|
||||
* Copyright (C) 2010, The Android Open Source Project
|
||||
*
|
||||
* Licensed under the Apache License, Version 2.0 (the "License");
|
||||
* you may not use this file except in compliance with the License.
|
||||
* You may obtain a copy of the License at
|
||||
*
|
||||
* http://www.apache.org/licenses/LICENSE-2.0
|
||||
*
|
||||
* Unless required by applicable law or agreed to in writing, software
|
||||
* distributed under the License is distributed on an "AS IS" BASIS,
|
||||
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
||||
* See the License for the specific language governing permissions and
|
||||
* limitations under the License.
|
||||
*/
|
||||
|
||||
#define LOG_TAG "LatinIME: bigram_dictionary.cpp"
|
||||
|
||||
#include "bigram_dictionary.h"
|
||||
|
||||
#include <algorithm>
|
||||
#include <cstring>
|
||||
|
||||
#include "defines.h"
|
||||
#include "suggest/core/dictionary/binary_dictionary_bigrams_iterator.h"
|
||||
#include "suggest/core/dictionary/dictionary.h"
|
||||
#include "suggest/core/policy/dictionary_structure_with_buffer_policy.h"
|
||||
#include "suggest/core/result/suggestion_results.h"
|
||||
#include "suggest/core/session/prev_words_info.h"
|
||||
#include "utils/char_utils.h"
|
||||
|
||||
namespace latinime {
|
||||
|
||||
BigramDictionary::BigramDictionary(
|
||||
const DictionaryStructureWithBufferPolicy *const dictionaryStructurePolicy)
|
||||
: mDictionaryStructurePolicy(dictionaryStructurePolicy) {
|
||||
if (DEBUG_DICT) {
|
||||
AKLOGI("BigramDictionary - constructor");
|
||||
}
|
||||
}
|
||||
|
||||
BigramDictionary::~BigramDictionary() {
|
||||
}
|
||||
|
||||
/* Parameters :
|
||||
* prevWordsInfo: Information of previous words to get the predictions.
|
||||
* outSuggestionResults: SuggestionResults to put the predictions.
|
||||
*/
|
||||
void BigramDictionary::getPredictions(const PrevWordsInfo *const prevWordsInfo,
|
||||
SuggestionResults *const outSuggestionResults) const {
|
||||
int unigramProbability = 0;
|
||||
int bigramCodePoints[MAX_WORD_LENGTH];
|
||||
BinaryDictionaryBigramsIterator bigramsIt =
|
||||
prevWordsInfo->getBigramsIteratorForPrediction(mDictionaryStructurePolicy);
|
||||
while (bigramsIt.hasNext()) {
|
||||
bigramsIt.next();
|
||||
if (bigramsIt.getBigramPos() == NOT_A_DICT_POS) {
|
||||
continue;
|
||||
}
|
||||
if (prevWordsInfo->isNthPrevWordBeginningOfSentence(1 /* n */)
|
||||
&& bigramsIt.getProbability() == NOT_A_PROBABILITY) {
|
||||
continue;
|
||||
}
|
||||
const int codePointCount = mDictionaryStructurePolicy->
|
||||
getCodePointsAndProbabilityAndReturnCodePointCount(bigramsIt.getBigramPos(),
|
||||
MAX_WORD_LENGTH, bigramCodePoints, &unigramProbability);
|
||||
if (codePointCount <= 0) {
|
||||
continue;
|
||||
}
|
||||
// Due to space constraints, the probability for bigrams is approximate - the lower the
|
||||
// unigram probability, the worse the precision. The theoritical maximum error in
|
||||
// resulting probability is 8 - although in the practice it's never bigger than 3 or 4
|
||||
// in very bad cases. This means that sometimes, we'll see some bigrams interverted
|
||||
// here, but it can't get too bad.
|
||||
const int probability = mDictionaryStructurePolicy->getProbability(
|
||||
unigramProbability, bigramsIt.getProbability());
|
||||
outSuggestionResults->addPrediction(bigramCodePoints, codePointCount, probability);
|
||||
}
|
||||
}
|
||||
|
||||
// Returns a pointer to the start of the bigram list.
|
||||
// If the word is not found or has no bigrams, this function returns NOT_A_DICT_POS.
|
||||
int BigramDictionary::getBigramListPositionForWord(const int *prevWord, const int prevWordLength,
|
||||
const bool forceLowerCaseSearch) const {
|
||||
if (0 >= prevWordLength) return NOT_A_DICT_POS;
|
||||
int pos = mDictionaryStructurePolicy->getTerminalPtNodePositionOfWord(prevWord, prevWordLength,
|
||||
forceLowerCaseSearch);
|
||||
if (NOT_A_DICT_POS == pos) return NOT_A_DICT_POS;
|
||||
return mDictionaryStructurePolicy->getBigramsPositionOfPtNode(pos);
|
||||
}
|
||||
|
||||
int BigramDictionary::getBigramProbability(const PrevWordsInfo *const prevWordsInfo,
|
||||
const int *word1, int length1) const {
|
||||
int nextWordPos = mDictionaryStructurePolicy->getTerminalPtNodePositionOfWord(word1, length1,
|
||||
false /* forceLowerCaseSearch */);
|
||||
if (NOT_A_DICT_POS == nextWordPos) return NOT_A_PROBABILITY;
|
||||
BinaryDictionaryBigramsIterator bigramsIt =
|
||||
prevWordsInfo->getBigramsIteratorForPrediction(mDictionaryStructurePolicy);
|
||||
while (bigramsIt.hasNext()) {
|
||||
bigramsIt.next();
|
||||
if (bigramsIt.getBigramPos() == nextWordPos
|
||||
&& bigramsIt.getProbability() != NOT_A_PROBABILITY) {
|
||||
return mDictionaryStructurePolicy->getProbability(
|
||||
mDictionaryStructurePolicy->getUnigramProbabilityOfPtNode(nextWordPos),
|
||||
bigramsIt.getProbability());
|
||||
}
|
||||
}
|
||||
return NOT_A_PROBABILITY;
|
||||
}
|
||||
|
||||
// TODO: Move functions related to bigram to here
|
||||
} // namespace latinime
|
@ -1,47 +0,0 @@
|
||||
/*
|
||||
* Copyright (C) 2010 The Android Open Source Project
|
||||
*
|
||||
* Licensed under the Apache License, Version 2.0 (the "License");
|
||||
* you may not use this file except in compliance with the License.
|
||||
* You may obtain a copy of the License at
|
||||
*
|
||||
* http://www.apache.org/licenses/LICENSE-2.0
|
||||
*
|
||||
* Unless required by applicable law or agreed to in writing, software
|
||||
* distributed under the License is distributed on an "AS IS" BASIS,
|
||||
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
||||
* See the License for the specific language governing permissions and
|
||||
* limitations under the License.
|
||||
*/
|
||||
|
||||
#ifndef LATINIME_BIGRAM_DICTIONARY_H
|
||||
#define LATINIME_BIGRAM_DICTIONARY_H
|
||||
|
||||
#include "defines.h"
|
||||
|
||||
namespace latinime {
|
||||
|
||||
class DictionaryStructureWithBufferPolicy;
|
||||
class PrevWordsInfo;
|
||||
class SuggestionResults;
|
||||
|
||||
class BigramDictionary {
|
||||
public:
|
||||
BigramDictionary(const DictionaryStructureWithBufferPolicy *const dictionaryStructurePolicy);
|
||||
|
||||
void getPredictions(const PrevWordsInfo *const prevWordsInfo,
|
||||
SuggestionResults *const outSuggestionResults) const;
|
||||
int getBigramProbability(const PrevWordsInfo *const prevWordsInfo,
|
||||
const int *word1, int length1) const;
|
||||
~BigramDictionary();
|
||||
|
||||
private:
|
||||
DISALLOW_IMPLICIT_CONSTRUCTORS(BigramDictionary);
|
||||
|
||||
int getBigramListPositionForWord(const int *prevWord, const int prevWordLength,
|
||||
const bool forceLowerCaseSearch) const;
|
||||
|
||||
const DictionaryStructureWithBufferPolicy *const mDictionaryStructurePolicy;
|
||||
};
|
||||
} // namespace latinime
|
||||
#endif // LATINIME_BIGRAM_DICTIONARY_H
|
@ -23,6 +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/suggest.h"
|
||||
#include "suggest/core/suggest_options.h"
|
||||
#include "suggest/policyimpl/gesture/gesture_suggest_policy_factory.h"
|
||||
@ -37,7 +38,6 @@ const int Dictionary::HEADER_ATTRIBUTE_BUFFER_SIZE = 32;
|
||||
Dictionary::Dictionary(JNIEnv *env, DictionaryStructureWithBufferPolicy::StructurePolicyPtr
|
||||
dictionaryStructureWithBufferPolicy)
|
||||
: mDictionaryStructureWithBufferPolicy(std::move(dictionaryStructureWithBufferPolicy)),
|
||||
mBigramDictionary(mDictionaryStructureWithBufferPolicy.get()),
|
||||
mGestureSuggest(new Suggest(GestureSuggestPolicyFactory::getGestureSuggestPolicy())),
|
||||
mTypingSuggest(new Suggest(TypingSuggestPolicyFactory::getTypingSuggestPolicy())) {
|
||||
logDictionaryInfo(env);
|
||||
@ -62,7 +62,29 @@ void Dictionary::getSuggestions(ProximityInfo *proximityInfo, DicTraverseSession
|
||||
void Dictionary::getPredictions(const PrevWordsInfo *const prevWordsInfo,
|
||||
SuggestionResults *const outSuggestionResults) const {
|
||||
TimeKeeper::setCurrentTime();
|
||||
mBigramDictionary.getPredictions(prevWordsInfo, outSuggestionResults);
|
||||
int unigramProbability = 0;
|
||||
int bigramCodePoints[MAX_WORD_LENGTH];
|
||||
BinaryDictionaryBigramsIterator bigramsIt = prevWordsInfo->getBigramsIteratorForPrediction(
|
||||
mDictionaryStructureWithBufferPolicy.get());
|
||||
while (bigramsIt.hasNext()) {
|
||||
bigramsIt.next();
|
||||
if (bigramsIt.getBigramPos() == NOT_A_DICT_POS) {
|
||||
continue;
|
||||
}
|
||||
if (prevWordsInfo->isNthPrevWordBeginningOfSentence(1 /* n */)
|
||||
&& bigramsIt.getProbability() == NOT_A_PROBABILITY) {
|
||||
continue;
|
||||
}
|
||||
const int codePointCount = mDictionaryStructureWithBufferPolicy->
|
||||
getCodePointsAndProbabilityAndReturnCodePointCount(bigramsIt.getBigramPos(),
|
||||
MAX_WORD_LENGTH, bigramCodePoints, &unigramProbability);
|
||||
if (codePointCount <= 0) {
|
||||
continue;
|
||||
}
|
||||
const int probability = mDictionaryStructureWithBufferPolicy->getProbability(
|
||||
unigramProbability, bigramsIt.getProbability());
|
||||
outSuggestionResults->addPrediction(bigramCodePoints, codePointCount, probability);
|
||||
}
|
||||
}
|
||||
|
||||
int Dictionary::getProbability(const int *word, int length) const {
|
||||
@ -84,7 +106,21 @@ int Dictionary::getMaxProbabilityOfExactMatches(const int *word, int length) con
|
||||
int Dictionary::getNgramProbability(const PrevWordsInfo *const prevWordsInfo, const int *word,
|
||||
int length) const {
|
||||
TimeKeeper::setCurrentTime();
|
||||
return mBigramDictionary.getBigramProbability(prevWordsInfo, word, length);
|
||||
int nextWordPos = mDictionaryStructureWithBufferPolicy->getTerminalPtNodePositionOfWord(word,
|
||||
length, false /* forceLowerCaseSearch */);
|
||||
if (NOT_A_DICT_POS == nextWordPos) return NOT_A_PROBABILITY;
|
||||
BinaryDictionaryBigramsIterator bigramsIt = prevWordsInfo->getBigramsIteratorForPrediction(
|
||||
mDictionaryStructureWithBufferPolicy.get());
|
||||
while (bigramsIt.hasNext()) {
|
||||
bigramsIt.next();
|
||||
if (bigramsIt.getBigramPos() == nextWordPos
|
||||
&& bigramsIt.getProbability() != NOT_A_PROBABILITY) {
|
||||
return mDictionaryStructureWithBufferPolicy->getProbability(
|
||||
mDictionaryStructureWithBufferPolicy->getUnigramProbabilityOfPtNode(
|
||||
nextWordPos), bigramsIt.getProbability());
|
||||
}
|
||||
}
|
||||
return NOT_A_PROBABILITY;
|
||||
}
|
||||
|
||||
bool Dictionary::addUnigramEntry(const int *const word, const int length,
|
||||
|
@ -21,7 +21,6 @@
|
||||
|
||||
#include "defines.h"
|
||||
#include "jni.h"
|
||||
#include "suggest/core/dictionary/bigram_dictionary.h"
|
||||
#include "suggest/core/dictionary/property/word_property.h"
|
||||
#include "suggest/core/policy/dictionary_header_structure_policy.h"
|
||||
#include "suggest/core/policy/dictionary_structure_with_buffer_policy.h"
|
||||
@ -119,7 +118,6 @@ class Dictionary {
|
||||
|
||||
const DictionaryStructureWithBufferPolicy::StructurePolicyPtr
|
||||
mDictionaryStructureWithBufferPolicy;
|
||||
const BigramDictionary mBigramDictionary;
|
||||
const SuggestInterfacePtr mGestureSuggest;
|
||||
const SuggestInterfacePtr mTypingSuggest;
|
||||
|
||||
|
@ -268,6 +268,10 @@ int PatriciaTriePolicy::getTerminalPtNodePositionOfWord(const int *const inWord,
|
||||
|
||||
int PatriciaTriePolicy::getProbability(const int unigramProbability,
|
||||
const int bigramProbability) const {
|
||||
// Due to space constraints, the probability for bigrams is approximate - the lower the unigram
|
||||
// probability, the worse the precision. The theoritical maximum error in resulting probability
|
||||
// is 8 - although in the practice it's never bigger than 3 or 4 in very bad cases. This means
|
||||
// that sometimes, we'll see some bigrams interverted here, but it can't get too bad.
|
||||
if (unigramProbability == NOT_A_PROBABILITY) {
|
||||
return NOT_A_PROBABILITY;
|
||||
} else if (bigramProbability == NOT_A_PROBABILITY) {
|
||||
|
Loading…
Reference in New Issue
Block a user