aboutsummaryrefslogtreecommitdiffstats
diff options
context:
space:
mode:
authorsatok <satok@google.com>2011-12-14 21:38:11 +0900
committersatok <satok@google.com>2011-12-15 12:09:25 +0900
commitd03317c4be21ee65c19d00c7b83a7042042b8627 (patch)
tree52c0685218832a4036e5cc4d1f2883feeca4297b
parent5a39e527dde07cc4360234553c9bb75694ece1ef (diff)
downloadlatinime-d03317c4be21ee65c19d00c7b83a7042042b8627.tar.gz
latinime-d03317c4be21ee65c19d00c7b83a7042042b8627.tar.xz
latinime-d03317c4be21ee65c19d00c7b83a7042042b8627.zip
Prune traversing a bit agressively and add a flag not to do auto completion
+1 1 -1 2 +2 0 -2 0 +3 0 -3 0 +4 6 -4 1 +5 4 -5 3 +6 3 -6 10 +7 7 -7 5 Before: Total 42936.28 (sum of others 42814.63) After: Total 40860.56 (sum of others 40733.92) Change-Id: I6a3d52f31ec181970083358280c3ebaca0a1f63e
-rw-r--r--native/src/correction.cpp7
-rw-r--r--native/src/correction.h4
-rw-r--r--native/src/unigram_dictionary.cpp9
3 files changed, 13 insertions, 7 deletions
diff --git a/native/src/correction.cpp b/native/src/correction.cpp
index 22ee75a24..364913f38 100644
--- a/native/src/correction.cpp
+++ b/native/src/correction.cpp
@@ -145,7 +145,7 @@ void Correction::initCorrectionState(
void Correction::setCorrectionParams(const int skipPos, const int excessivePos,
const int transposedPos, const int spaceProximityPos, const int missingSpacePos,
- const bool useFullEditDistance) {
+ const bool useFullEditDistance, const bool doAutoCompletion) {
// TODO: remove
mTransposedPos = transposedPos;
mExcessivePos = excessivePos;
@@ -158,6 +158,7 @@ void Correction::setCorrectionParams(const int skipPos, const int excessivePos,
mSpaceProximityPos = spaceProximityPos;
mMissingSpacePos = missingSpacePos;
mUseFullEditDistance = useFullEditDistance;
+ mDoAutoCompletion = doAutoCompletion;
}
void Correction::checkState() {
@@ -279,7 +280,9 @@ void Correction::startToTraverseAllNodes() {
bool Correction::needsToPrune() const {
// TODO: use edit distance here
- return mOutputIndex - 1 >= mMaxDepth || mProximityCount > mMaxEditDistance;
+ return mOutputIndex - 1 >= mMaxDepth || mProximityCount > mMaxEditDistance
+ // Allow one char longer word for missing character
+ || (!mDoAutoCompletion && (mOutputIndex + 1 >= mInputLength));
}
void Correction::addCharToCurrentWord(const int32_t c) {
diff --git a/native/src/correction.h b/native/src/correction.h
index d4e99f0ce..4a8d1fab7 100644
--- a/native/src/correction.h
+++ b/native/src/correction.h
@@ -44,7 +44,8 @@ public:
// TODO: remove
void setCorrectionParams(const int skipPos, const int excessivePos, const int transposedPos,
- const int spaceProximityPos, const int missingSpacePos, const bool useFullEditDistance);
+ const int spaceProximityPos, const int missingSpacePos, const bool useFullEditDistance,
+ const bool doAutoCompletion);
void checkState();
bool initProcessState(const int index);
@@ -109,6 +110,7 @@ private:
const ProximityInfo *mProximityInfo;
bool mUseFullEditDistance;
+ bool mDoAutoCompletion;
int mMaxEditDistance;
int mMaxDepth;
int mInputLength;
diff --git a/native/src/unigram_dictionary.cpp b/native/src/unigram_dictionary.cpp
index a2c1f72a1..7cf191970 100644
--- a/native/src/unigram_dictionary.cpp
+++ b/native/src/unigram_dictionary.cpp
@@ -260,7 +260,8 @@ void UnigramDictionary::getSuggestionCandidates(const bool useFullEditDistance,
const int inputLength, Correction *correction, WordsPriorityQueue *queue) {
// TODO: Remove setCorrectionParams
correction->setCorrectionParams(0, 0, 0,
- -1 /* spaceProximityPos */, -1 /* missingSpacePos */, useFullEditDistance);
+ -1 /* spaceProximityPos */, -1 /* missingSpacePos */, useFullEditDistance,
+ true /* doAutoCompletion */);
int rootPosition = ROOT_POS;
// Get the number of children of root, then increment the position
int childCount = Dictionary::getCount(DICT_ROOT, &rootPosition);
@@ -295,7 +296,7 @@ void UnigramDictionary::getMissingSpaceWords(
Correction *correction, const bool useFullEditDistance, WordsPriorityQueue *queue) {
correction->setCorrectionParams(-1 /* skipPos */, -1 /* excessivePos */,
-1 /* transposedPos */, -1 /* spaceProximityPos */, missingSpacePos,
- useFullEditDistance);
+ useFullEditDistance, true /* doAutoCompletion */);
getSplitTwoWordsSuggestion(inputLength, proximityInfo, correction, queue);
}
@@ -304,7 +305,7 @@ void UnigramDictionary::getMistypedSpaceWords(
Correction *correction, const bool useFullEditDistance, WordsPriorityQueue *queue) {
correction->setCorrectionParams(-1 /* skipPos */, -1 /* excessivePos */,
-1 /* transposedPos */, spaceProximityPos, -1 /* missingSpacePos */,
- useFullEditDistance);
+ useFullEditDistance, true /* doAutoCompletion */);
getSplitTwoWordsSuggestion(inputLength, proximityInfo, correction, queue);
}
@@ -585,7 +586,7 @@ inline bool UnigramDictionary::processCurrentNode(const int initialPos,
if (stateType == Correction::TRAVERSE_ALL_ON_TERMINAL
|| stateType == Correction::ON_TERMINAL) {
needsToInvokeOnTerminal = true;
- } else if (stateType == Correction::UNRELATED) {
+ } else if (stateType == Correction::UNRELATED || correction->needsToPrune()) {
// We found that this is an unrelated character, so we should give up traversing
// this node and its children entirely.
// However we may not be on the last virtual node yet so we skip the remaining