Merge NOT_A_VALID_WORD_POS into NOT_A_DICT_POS.

Change-Id: I2ac3aa2a19d14e8c4159a59334e7816bd1409688
main
Keisuke Kuroyanagi 2013-09-16 18:30:35 +09:00
parent 9ea9c61c99
commit cb816e5e16
17 changed files with 50 additions and 52 deletions

View File

@ -292,7 +292,6 @@ static inline void prof_out(void) {
// of the binary dictionary where a {key,value} string pair scheme is used.
#define LARGEST_INT_DIGIT_COUNT 11
#define NOT_A_VALID_WORD_POS (-99)
#define NOT_A_CODE_POINT (-1)
#define NOT_A_DISTANCE (-1)
#define NOT_A_COORDINATE (-1)

View File

@ -112,7 +112,7 @@ class DicNode {
mIsUsed = true;
mIsCachedForNextSuggestion = false;
mDicNodeProperties.init(
NOT_A_VALID_WORD_POS /* pos */, rootGroupPos, NOT_A_CODE_POINT /* nodeCodePoint */,
NOT_A_DICT_POS /* pos */, rootGroupPos, NOT_A_CODE_POINT /* nodeCodePoint */,
NOT_A_PROBABILITY /* probability */, false /* isTerminal */,
true /* hasChildren */, false /* isBlacklistedOrNotAWord */, 0 /* depth */,
0 /* terminalDepth */);
@ -125,7 +125,7 @@ class DicNode {
mIsUsed = true;
mIsCachedForNextSuggestion = dicNode->mIsCachedForNextSuggestion;
mDicNodeProperties.init(
NOT_A_VALID_WORD_POS /* pos */, rootGroupPos, NOT_A_CODE_POINT /* nodeCodePoint */,
NOT_A_DICT_POS /* pos */, rootGroupPos, NOT_A_CODE_POINT /* nodeCodePoint */,
NOT_A_PROBABILITY /* probability */, false /* isTerminal */,
true /* hasChildren */, false /* isBlacklistedOrNotAWord */, 0 /* depth */,
0 /* terminalDepth */);
@ -234,7 +234,7 @@ class DicNode {
}
bool isFirstWord() const {
return mDicNodeState.mDicNodeStatePrevWord.getPrevWordNodePos() == NOT_A_VALID_WORD_POS;
return mDicNodeState.mDicNodeStatePrevWord.getPrevWordNodePos() == NOT_A_DICT_POS;
}
bool isCompletion(const int inputSize) const {

View File

@ -89,7 +89,7 @@ namespace latinime {
const int unigramProbability = node->getProbability();
const int wordPos = node->getPos();
const int prevWordPos = node->getPrevWordPos();
if (NOT_A_VALID_WORD_POS == wordPos || NOT_A_VALID_WORD_POS == prevWordPos) {
if (NOT_A_DICT_POS == wordPos || NOT_A_DICT_POS == prevWordPos) {
// Note: Normally wordPos comes from the dictionary and should never equal
// NOT_A_VALID_WORD_POS.
return dictionaryStructurePolicy->getProbability(unigramProbability,

View File

@ -29,7 +29,7 @@ class DicNodeStatePrevWord {
public:
AK_FORCE_INLINE DicNodeStatePrevWord()
: mPrevWordCount(0), mPrevWordLength(0), mPrevWordStart(0), mPrevWordProbability(0),
mPrevWordNodePos(NOT_A_VALID_WORD_POS) {
mPrevWordNodePos(NOT_A_DICT_POS) {
memset(mPrevWord, 0, sizeof(mPrevWord));
memset(mPrevSpacePositions, 0, sizeof(mPrevSpacePositions));
}
@ -41,7 +41,7 @@ class DicNodeStatePrevWord {
mPrevWordCount = 0;
mPrevWordStart = 0;
mPrevWordProbability = -1;
mPrevWordNodePos = NOT_A_VALID_WORD_POS;
mPrevWordNodePos = NOT_A_DICT_POS;
memset(mPrevSpacePositions, 0, sizeof(mPrevSpacePositions));
}

View File

@ -116,7 +116,7 @@ int BigramDictionary::getPredictions(const int *prevWord, const int prevWordLeng
mDictionaryStructurePolicy->getBigramsStructurePolicy(), pos);
while (bigramsIt.hasNext()) {
bigramsIt.next();
if (bigramsIt.getBigramPos() == NOT_A_VALID_WORD_POS) {
if (bigramsIt.getBigramPos() == NOT_A_DICT_POS) {
continue;
}
const int codePointCount = mDictionaryStructurePolicy->
@ -146,7 +146,7 @@ int BigramDictionary::getBigramListPositionForWord(const int *prevWord, const in
if (0 >= prevWordLength) return NOT_A_DICT_POS;
int pos = mDictionaryStructurePolicy->getTerminalNodePositionOfWord(prevWord, prevWordLength,
forceLowerCaseSearch);
if (NOT_A_VALID_WORD_POS == pos) return NOT_A_DICT_POS;
if (NOT_A_DICT_POS == pos) return NOT_A_DICT_POS;
return mDictionaryStructurePolicy->getBigramsPositionOfNode(pos);
}
@ -157,7 +157,7 @@ bool BigramDictionary::isValidBigram(const int *word0, int length0, const int *w
if (NOT_A_DICT_POS == pos) return false;
int nextWordPos = mDictionaryStructurePolicy->getTerminalNodePositionOfWord(word1, length1,
false /* forceLowerCaseSearch */);
if (NOT_A_VALID_WORD_POS == nextWordPos) return false;
if (NOT_A_DICT_POS == nextWordPos) return false;
BinaryDictionaryBigramsIterator bigramsIt(
mDictionaryStructurePolicy->getBigramsStructurePolicy(), pos);

View File

@ -87,7 +87,7 @@ int Dictionary::getBigrams(const int *word, int length, int *outWords, int *freq
int Dictionary::getProbability(const int *word, int length) const {
int pos = getDictionaryStructurePolicy()->getTerminalNodePositionOfWord(word, length,
false /* forceLowerCaseSearch */);
if (NOT_A_VALID_WORD_POS == pos) {
if (NOT_A_DICT_POS == pos) {
return NOT_A_PROBABILITY;
}
return getDictionaryStructurePolicy()->getUnigramProbabilityOfPtNode(pos);

View File

@ -73,7 +73,7 @@ class MultiBigramMap {
bigramsListPos);
while (bigramsIt.hasNext()) {
bigramsIt.next();
if (bigramsIt.getBigramPos() == NOT_A_VALID_WORD_POS) {
if (bigramsIt.getBigramPos() == NOT_A_DICT_POS) {
continue;
}
mBigramMap[bigramsIt.getBigramPos()] = bigramsIt.getProbability();

View File

@ -35,13 +35,13 @@ void DicTraverseSession::init(const Dictionary *const dictionary, const int *pre
->getMultiWordCostMultiplier();
mSuggestOptions = suggestOptions;
if (!prevWord) {
mPrevWordPos = NOT_A_VALID_WORD_POS;
mPrevWordPos = NOT_A_DICT_POS;
return;
}
// TODO: merge following similar calls to getTerminalPosition into one case-insensitive call.
mPrevWordPos = getDictionaryStructurePolicy()->getTerminalNodePositionOfWord(
prevWord, prevWordLength, false /* forceLowerCaseSearch */);
if (mPrevWordPos == NOT_A_VALID_WORD_POS) {
if (mPrevWordPos == NOT_A_DICT_POS) {
// Check bigrams for lower-cased previous word if original was not found. Useful for
// auto-capitalized words like "The [current_word]".
mPrevWordPos = getDictionaryStructurePolicy()->getTerminalNodePositionOfWord(

View File

@ -59,7 +59,7 @@ class DicTraverseSession {
}
AK_FORCE_INLINE DicTraverseSession(JNIEnv *env, jstring localeStr, bool usesLargeCache)
: mPrevWordPos(NOT_A_VALID_WORD_POS), mProximityInfo(0),
: mPrevWordPos(NOT_A_DICT_POS), mProximityInfo(0),
mDictionary(0), mSuggestOptions(0), mDicNodesCache(usesLargeCache),
mMultiBigramMap(), mInputSize(0), mPartiallyCommited(false), mMaxPointerCount(1),
mMultiWordCostMultiplier(1.0f) {

View File

@ -70,7 +70,7 @@ const int BigramListReadWriteUtils::ATTRIBUTE_ADDRESS_SHIFT = 4;
break;
}
if (offset == 0) {
return NOT_A_VALID_WORD_POS;
return NOT_A_DICT_POS;
}
if (isOffsetNegative(flags)) {
return origin - offset;

View File

@ -67,7 +67,7 @@ public:
const int entryPos, const int targetPos, const int probability, const bool hasNext,
BigramFlags *const outBigramFlags, uint32_t *const outOffset,
int *const outOffsetFieldSize) {
if (targetPos == NOT_A_VALID_WORD_POS) {
if (targetPos == NOT_A_DICT_POS) {
return false;
}
BigramFlags flags = probability & MASK_ATTRIBUTE_PROBABILITY;

View File

@ -31,7 +31,7 @@ void DynamicBigramListPolicy::getNextBigram(int *const outBigramPos, int *const
BigramListReadWriteUtils::getFlagsAndForwardPointer(buffer, pos);
int originalBigramPos = BigramListReadWriteUtils::getBigramAddressAndForwardPointer(
buffer, flags, pos);
if (usesAdditionalBuffer && originalBigramPos != NOT_A_VALID_WORD_POS) {
if (usesAdditionalBuffer && originalBigramPos != NOT_A_DICT_POS) {
originalBigramPos += mBuffer->getOriginalBufferSize();
}
*outBigramPos = followBigramLinkAndGetCurrentBigramPtNodePos(originalBigramPos);
@ -66,7 +66,7 @@ bool DynamicBigramListPolicy::copyAllBigrams(int *const fromPos, int *const toPo
flags = BigramListReadWriteUtils::getFlagsAndForwardPointer(buffer, fromPos);
int originalBigramPos = BigramListReadWriteUtils::getBigramAddressAndForwardPointer(
buffer, flags, fromPos);
if (originalBigramPos == NOT_A_VALID_WORD_POS) {
if (originalBigramPos == NOT_A_DICT_POS) {
// skip invalid bigram entry.
continue;
}
@ -172,7 +172,7 @@ bool DynamicBigramListPolicy::removeBigram(const int bigramListPos, const int ta
}
int originalBigramPos = BigramListReadWriteUtils::getBigramAddressAndForwardPointer(
buffer, flags, &pos);
if (usesAdditionalBuffer && originalBigramPos != NOT_A_VALID_WORD_POS) {
if (usesAdditionalBuffer && originalBigramPos != NOT_A_DICT_POS) {
originalBigramPos += mBuffer->getOriginalBufferSize();
}
const int bigramPos = followBigramLinkAndGetCurrentBigramPtNodePos(originalBigramPos);
@ -192,8 +192,8 @@ bool DynamicBigramListPolicy::removeBigram(const int bigramListPos, const int ta
int DynamicBigramListPolicy::followBigramLinkAndGetCurrentBigramPtNodePos(
const int originalBigramPos) const {
if (originalBigramPos == NOT_A_VALID_WORD_POS) {
return NOT_A_VALID_WORD_POS;
if (originalBigramPos == NOT_A_DICT_POS) {
return NOT_A_DICT_POS;
}
int currentPos = originalBigramPos;
DynamicPatriciaTrieNodeReader nodeReader(mBuffer, this /* bigramsPolicy */, mShortcutPolicy);
@ -206,7 +206,7 @@ int DynamicBigramListPolicy::followBigramLinkAndGetCurrentBigramPtNodePos(
if (bigramLinkCount > BIGRAM_LINK_COUNT_LIMIT) {
AKLOGI("Bigram link is invalid. start position: %d", bigramPos);
ASSERT(false);
return NOT_A_VALID_WORD_POS;
return NOT_A_DICT_POS;
}
}
return currentPos;

View File

@ -69,7 +69,7 @@ void DynamicPatriciaTrieNodeReader::fetchNodeInfoFromBufferAndProcessMovedNode(c
if (usesAdditionalBuffer && mChildrenPos != NOT_A_DICT_POS) {
mChildrenPos += mBuffer->getOriginalBufferSize();
}
if (mSiblingPos == NOT_A_VALID_WORD_POS && DynamicPatriciaTrieReadingUtils::isMoved(mFlags)) {
if (mSiblingPos == NOT_A_DICT_POS && DynamicPatriciaTrieReadingUtils::isMoved(mFlags)) {
mBigramLinkedNodePos = mChildrenPos;
} else {
mBigramLinkedNodePos = NOT_A_DICT_POS;
@ -90,7 +90,7 @@ void DynamicPatriciaTrieNodeReader::fetchNodeInfoFromBufferAndProcessMovedNode(c
mBigramPos = NOT_A_DICT_POS;
}
// Update siblingPos if needed.
if (mSiblingPos == NOT_A_VALID_WORD_POS) {
if (mSiblingPos == NOT_A_DICT_POS) {
// Sibling position is the tail position of current node.
mSiblingPos = pos;
}
@ -102,7 +102,7 @@ void DynamicPatriciaTrieNodeReader::fetchNodeInfoFromBufferAndProcessMovedNode(c
}
void DynamicPatriciaTrieNodeReader::invalidatePtNodeInfo() {
mHeadPos = NOT_A_VALID_WORD_POS;
mHeadPos = NOT_A_DICT_POS;
mFlags = 0;
mParentPos = NOT_A_DICT_POS;
mCodePointCount = 0;
@ -113,7 +113,7 @@ void DynamicPatriciaTrieNodeReader::invalidatePtNodeInfo() {
mBigramLinkedNodePos = NOT_A_DICT_POS;
mShortcutPos = NOT_A_DICT_POS;
mBigramPos = NOT_A_DICT_POS;
mSiblingPos = NOT_A_VALID_WORD_POS;
mSiblingPos = NOT_A_DICT_POS;
}
}

View File

@ -39,12 +39,12 @@ class DynamicPatriciaTrieNodeReader {
const DictionaryBigramsStructurePolicy *const bigramsPolicy,
const DictionaryShortcutsStructurePolicy *const shortcutsPolicy)
: mBuffer(buffer), mBigramsPolicy(bigramsPolicy),
mShortcutsPolicy(shortcutsPolicy), mHeadPos(NOT_A_VALID_WORD_POS), mFlags(0),
mShortcutsPolicy(shortcutsPolicy), mHeadPos(NOT_A_DICT_POS), mFlags(0),
mParentPos(NOT_A_DICT_POS), mCodePointCount(0), mProbabilityFieldPos(NOT_A_DICT_POS),
mProbability(NOT_A_PROBABILITY), mChildrenPosFieldPos(NOT_A_DICT_POS),
mChildrenPos(NOT_A_DICT_POS), mBigramLinkedNodePos(NOT_A_DICT_POS),
mShortcutPos(NOT_A_DICT_POS), mBigramPos(NOT_A_DICT_POS),
mSiblingPos(NOT_A_VALID_WORD_POS) {}
mSiblingPos(NOT_A_DICT_POS) {}
~DynamicPatriciaTrieNodeReader() {}
@ -56,7 +56,7 @@ class DynamicPatriciaTrieNodeReader {
AK_FORCE_INLINE void fetchNodeInfoFromBufferAndGetNodeCodePoints(const int nodePos,
const int maxCodePointCount, int *const outCodePoints) {
mSiblingPos = NOT_A_VALID_WORD_POS;
mSiblingPos = NOT_A_DICT_POS;
mBigramLinkedNodePos = NOT_A_DICT_POS;
fetchNodeInfoFromBufferAndProcessMovedNode(nodePos, maxCodePointCount, outCodePoints);
}

View File

@ -116,7 +116,7 @@ int DynamicPatriciaTriePolicy::getTerminalNodePositionOfWord(const int *const in
if (!readingHelper.isMatchedCodePoint(
j, searchCodePoints[matchedCodePointCount + j])) {
// Different code point is found. The given word is not included in the dictionary.
return NOT_A_VALID_WORD_POS;
return NOT_A_DICT_POS;
}
}
// All characters are matched.
@ -125,14 +125,14 @@ int DynamicPatriciaTriePolicy::getTerminalNodePositionOfWord(const int *const in
return nodeReader->getHeadPos();
}
if (!nodeReader->hasChildren()) {
return NOT_A_VALID_WORD_POS;
return NOT_A_DICT_POS;
}
// Advance to the children nodes.
readingHelper.readChildNode();
}
// If we already traversed the tree further than the word is long, there means
// there was no match (or we would have found it).
return NOT_A_VALID_WORD_POS;
return NOT_A_DICT_POS;
}
int DynamicPatriciaTriePolicy::getProbability(const int unigramProbability,
@ -149,7 +149,7 @@ int DynamicPatriciaTriePolicy::getProbability(const int unigramProbability,
}
int DynamicPatriciaTriePolicy::getUnigramProbabilityOfPtNode(const int nodePos) const {
if (nodePos == NOT_A_VALID_WORD_POS) {
if (nodePos == NOT_A_DICT_POS) {
return NOT_A_PROBABILITY;
}
DynamicPatriciaTrieNodeReader nodeReader(&mBufferWithExtendableBuffer,
@ -162,7 +162,7 @@ int DynamicPatriciaTriePolicy::getUnigramProbabilityOfPtNode(const int nodePos)
}
int DynamicPatriciaTriePolicy::getShortcutPositionOfNode(const int nodePos) const {
if (nodePos == NOT_A_VALID_WORD_POS) {
if (nodePos == NOT_A_DICT_POS) {
return NOT_A_DICT_POS;
}
DynamicPatriciaTrieNodeReader nodeReader(&mBufferWithExtendableBuffer,
@ -175,7 +175,7 @@ int DynamicPatriciaTriePolicy::getShortcutPositionOfNode(const int nodePos) cons
}
int DynamicPatriciaTriePolicy::getBigramsPositionOfNode(const int nodePos) const {
if (nodePos == NOT_A_VALID_WORD_POS) {
if (nodePos == NOT_A_DICT_POS) {
return NOT_A_DICT_POS;
}
DynamicPatriciaTrieNodeReader nodeReader(&mBufferWithExtendableBuffer,
@ -209,12 +209,12 @@ bool DynamicPatriciaTriePolicy::addBigramWords(const int *const word0, const int
}
const int word0Pos = getTerminalNodePositionOfWord(word0, length0,
false /* forceLowerCaseSearch */);
if (word0Pos == NOT_A_VALID_WORD_POS) {
if (word0Pos == NOT_A_DICT_POS) {
return false;
}
const int word1Pos = getTerminalNodePositionOfWord(word1, length1,
false /* forceLowerCaseSearch */);
if (word1Pos == NOT_A_VALID_WORD_POS) {
if (word1Pos == NOT_A_DICT_POS) {
return false;
}
DynamicPatriciaTrieWritingHelper writingHelper(&mBufferWithExtendableBuffer,
@ -230,12 +230,12 @@ bool DynamicPatriciaTriePolicy::removeBigramWords(const int *const word0, const
}
const int word0Pos = getTerminalNodePositionOfWord(word0, length0,
false /* forceLowerCaseSearch */);
if (word0Pos == NOT_A_VALID_WORD_POS) {
if (word0Pos == NOT_A_DICT_POS) {
return false;
}
const int word1Pos = getTerminalNodePositionOfWord(word1, length1,
false /* forceLowerCaseSearch */);
if (word1Pos == NOT_A_VALID_WORD_POS) {
if (word1Pos == NOT_A_DICT_POS) {
return false;
}
DynamicPatriciaTrieWritingHelper writingHelper(&mBufferWithExtendableBuffer,

View File

@ -72,8 +72,7 @@ class DynamicPatriciaTrieReadingHelper {
// Initialize reading state with the head position of a node.
AK_FORCE_INLINE void initWithNodePos(const int nodePos) {
// TODO: Consolidate NOT_A_VALID_WORD_POS and NOT_A_DICT_POS
if (nodePos == NOT_A_VALID_WORD_POS || nodePos == NOT_A_DICT_POS) {
if (nodePos == NOT_A_DICT_POS) {
mPos = NOT_A_DICT_POS;
} else {
mIsError = false;

View File

@ -219,7 +219,7 @@ int PatriciaTriePolicy::getCodePointsAndProbabilityAndReturnCodePointCount(
}
// This function gets the position of the terminal node of the exact matching word in the
// dictionary. If no match is found, it returns NOT_A_VALID_WORD_POS.
// dictionary. If no match is found, it returns NOT_A_DICT_POS.
int PatriciaTriePolicy::getTerminalNodePositionOfWord(const int *const inWord,
const int length, const bool forceLowerCaseSearch) const {
int pos = getRootPosition();
@ -228,7 +228,7 @@ int PatriciaTriePolicy::getTerminalNodePositionOfWord(const int *const inWord,
while (true) {
// If we already traversed the tree further than the word is long, there means
// there was no match (or we would have found it).
if (wordPos >= length) return NOT_A_VALID_WORD_POS;
if (wordPos >= length) return NOT_A_DICT_POS;
int ptNodeCount = PatriciaTrieReadingUtils::getPtNodeArraySizeAndAdvancePosition(mDictRoot,
&pos);
const int wChar = forceLowerCaseSearch
@ -236,7 +236,7 @@ int PatriciaTriePolicy::getTerminalNodePositionOfWord(const int *const inWord,
while (true) {
// If there are no more PtNodes in this array, it means we could not
// find a matching character for this depth, therefore there is no match.
if (0 >= ptNodeCount) return NOT_A_VALID_WORD_POS;
if (0 >= ptNodeCount) return NOT_A_DICT_POS;
const int ptNodePos = pos;
const PatriciaTrieReadingUtils::NodeFlags flags =
PatriciaTrieReadingUtils::getFlagsAndAdvancePosition(mDictRoot, &pos);
@ -245,7 +245,7 @@ int PatriciaTriePolicy::getTerminalNodePositionOfWord(const int *const inWord,
if (character == wChar) {
// This is the correct PtNode. Only one PtNode may start with the same char within
// a PtNode array, so either we found our match in this array, or there is
// no match and we can return NOT_A_VALID_WORD_POS. So we will check all the
// no match and we can return NOT_A_DICT_POS. So we will check all the
// characters in this PtNode indeed does match.
if (PatriciaTrieReadingUtils::hasMultipleChars(flags)) {
character = PatriciaTrieReadingUtils::getCodePointAndAdvancePosition(mDictRoot,
@ -256,8 +256,8 @@ int PatriciaTriePolicy::getTerminalNodePositionOfWord(const int *const inWord,
// character that does not match, as explained above, it means the word is
// not in the dictionary (by virtue of this PtNode being the only one to
// match the word on the first character, but not matching the whole word).
if (wordPos >= length) return NOT_A_VALID_WORD_POS;
if (inWord[wordPos] != character) return NOT_A_VALID_WORD_POS;
if (wordPos >= length) return NOT_A_DICT_POS;
if (inWord[wordPos] != character) return NOT_A_DICT_POS;
character = PatriciaTrieReadingUtils::getCodePointAndAdvancePosition(
mDictRoot, &pos);
}
@ -274,7 +274,7 @@ int PatriciaTriePolicy::getTerminalNodePositionOfWord(const int *const inWord,
PatriciaTrieReadingUtils::readProbabilityAndAdvancePosition(mDictRoot, &pos);
}
if (!PatriciaTrieReadingUtils::hasChildrenInFlags(flags)) {
return NOT_A_VALID_WORD_POS;
return NOT_A_DICT_POS;
}
// We have children and we are still shorter than the word we are searching for, so
// we need to traverse children. Put the pointer on the children position, and
@ -320,7 +320,7 @@ int PatriciaTriePolicy::getProbability(const int unigramProbability,
}
int PatriciaTriePolicy::getUnigramProbabilityOfPtNode(const int nodePos) const {
if (nodePos == NOT_A_VALID_WORD_POS) {
if (nodePos == NOT_A_DICT_POS) {
return NOT_A_PROBABILITY;
}
int pos = nodePos;
@ -342,7 +342,7 @@ int PatriciaTriePolicy::getUnigramProbabilityOfPtNode(const int nodePos) const {
}
int PatriciaTriePolicy::getShortcutPositionOfNode(const int nodePos) const {
if (nodePos == NOT_A_VALID_WORD_POS) {
if (nodePos == NOT_A_DICT_POS) {
return NOT_A_DICT_POS;
}
int pos = nodePos;
@ -362,7 +362,7 @@ int PatriciaTriePolicy::getShortcutPositionOfNode(const int nodePos) const {
}
int PatriciaTriePolicy::getBigramsPositionOfNode(const int nodePos) const {
if (nodePos == NOT_A_VALID_WORD_POS) {
if (nodePos == NOT_A_DICT_POS) {
return NOT_A_DICT_POS;
}
int pos = nodePos;