Pass words as int[] to the native code.
We need to get the bigrams during the call to getSuggestions for
bug#6313806. We already give an int[] to getSuggestions and we
wanted to get rid of char[]'s anyway because it doesn't work with
surrogate pairs, so here we go.
Bug: 6313806
Change-Id: I56ce99f1db6b3302cdf42f0527343bded837091e
diff --git a/native/jni/com_android_inputmethod_latin_BinaryDictionary.cpp b/native/jni/com_android_inputmethod_latin_BinaryDictionary.cpp
index 2ef72e1..3e72ce6 100644
--- a/native/jni/com_android_inputmethod_latin_BinaryDictionary.cpp
+++ b/native/jni/com_android_inputmethod_latin_BinaryDictionary.cpp
@@ -153,30 +153,30 @@
}
static int latinime_BinaryDictionary_getBigrams(JNIEnv *env, jobject object, jlong dict,
- jcharArray prevWordArray, jint prevWordLength, jintArray inputArray, jint inputArraySize,
+ jintArray prevWordArray, jint prevWordLength, jintArray inputArray, jint inputArraySize,
jcharArray outputArray, jintArray frequencyArray, jint maxWordLength, jint maxBigrams) {
Dictionary *dictionary = (Dictionary*)dict;
if (!dictionary) return 0;
- jchar *prevWord = env->GetCharArrayElements(prevWordArray, 0);
+ jint *prevWord = env->GetIntArrayElements(prevWordArray, 0);
int *inputCodes = env->GetIntArrayElements(inputArray, 0);
jchar *outputChars = env->GetCharArrayElements(outputArray, 0);
int *frequencies = env->GetIntArrayElements(frequencyArray, 0);
- int count = dictionary->getBigrams((unsigned short*) prevWord, prevWordLength, inputCodes,
+ int count = dictionary->getBigrams(prevWord, prevWordLength, inputCodes,
inputArraySize, (unsigned short*) outputChars, frequencies, maxWordLength, maxBigrams);
env->ReleaseIntArrayElements(frequencyArray, frequencies, 0);
env->ReleaseCharArrayElements(outputArray, outputChars, 0);
env->ReleaseIntArrayElements(inputArray, inputCodes, JNI_ABORT);
- env->ReleaseCharArrayElements(prevWordArray, prevWord, JNI_ABORT);
+ env->ReleaseIntArrayElements(prevWordArray, prevWord, JNI_ABORT);
return count;
}
static jboolean latinime_BinaryDictionary_isValidWord(JNIEnv *env, jobject object, jlong dict,
- jcharArray wordArray, jint wordLength) {
+ jintArray wordArray, jint wordLength) {
Dictionary *dictionary = (Dictionary*)dict;
if (!dictionary) return (jboolean) false;
- jchar *word = env->GetCharArrayElements(wordArray, 0);
- jboolean result = dictionary->isValidWord((unsigned short*) word, wordLength);
- env->ReleaseCharArrayElements(wordArray, word, JNI_ABORT);
+ jint *word = env->GetIntArrayElements(wordArray, 0);
+ jboolean result = dictionary->isValidWord(word, wordLength);
+ env->ReleaseIntArrayElements(wordArray, word, JNI_ABORT);
return result;
}
@@ -236,8 +236,8 @@
{"closeNative", "(J)V", (void*)latinime_BinaryDictionary_close},
{"getSuggestionsNative", "(JJ[I[I[II[IZ[C[I)I",
(void*)latinime_BinaryDictionary_getSuggestions},
- {"isValidWordNative", "(J[CI)Z", (void*)latinime_BinaryDictionary_isValidWord},
- {"getBigramsNative", "(J[CI[II[C[III)I", (void*)latinime_BinaryDictionary_getBigrams},
+ {"isValidWordNative", "(J[II)Z", (void*)latinime_BinaryDictionary_isValidWord},
+ {"getBigramsNative", "(J[II[II[C[III)I", (void*)latinime_BinaryDictionary_getBigrams},
{"calcNormalizedScoreNative", "([CI[CII)D",
(void*)latinime_BinaryDictionary_calcNormalizedScore},
{"editDistanceNative", "([CI[CI)I", (void*)latinime_BinaryDictionary_editDistance}
diff --git a/native/jni/src/bigram_dictionary.cpp b/native/jni/src/bigram_dictionary.cpp
index 320b0af..927381f 100644
--- a/native/jni/src/bigram_dictionary.cpp
+++ b/native/jni/src/bigram_dictionary.cpp
@@ -96,7 +96,7 @@
* and the bigrams are used to boost unigram result scores, it makes little sense to
* reduce their scope to the ones that match the first letter.
*/
-int BigramDictionary::getBigrams(unsigned short *prevWord, int prevWordLength, int *codes,
+int BigramDictionary::getBigrams(const int32_t *prevWord, int prevWordLength, int *codes,
int codesSize, unsigned short *bigramChars, int *bigramFreq, int maxWordLength,
int maxBigrams) {
// TODO: remove unused arguments, and refrain from storing stuff in members of this class
@@ -134,7 +134,7 @@
// Returns a pointer to the start of the bigram list.
// If the word is not found or has no bigrams, this function returns 0.
int BigramDictionary::getBigramListForWord(const uint8_t* const root,
- const unsigned short *prevWord, const int prevWordLength) {
+ const int32_t *prevWord, const int prevWordLength) {
int pos = BinaryFormat::getTerminalPosition(root, prevWord, prevWordLength);
if (NOT_VALID_WORD == pos) return 0;
diff --git a/native/jni/src/bigram_dictionary.h b/native/jni/src/bigram_dictionary.h
index 1612131..07e47f0 100644
--- a/native/jni/src/bigram_dictionary.h
+++ b/native/jni/src/bigram_dictionary.h
@@ -25,10 +25,10 @@
class BigramDictionary {
public:
BigramDictionary(const unsigned char *dict, int maxWordLength, Dictionary *parentDictionary);
- int getBigrams(unsigned short *word, int length, int *codes, int codesSize,
+ int getBigrams(const int32_t *word, int length, int *codes, int codesSize,
unsigned short *outWords, int *frequencies, int maxWordLength, int maxBigrams);
int getBigramListForWord(const uint8_t* const root,
- const unsigned short *prevWord, const int prevWordLength);
+ const int32_t *prevWord, const int prevWordLength);
~BigramDictionary();
private:
bool addWordBigram(unsigned short *word, int length, int frequency);
diff --git a/native/jni/src/binary_format.h b/native/jni/src/binary_format.h
index f593024..b8ac952 100644
--- a/native/jni/src/binary_format.h
+++ b/native/jni/src/binary_format.h
@@ -62,7 +62,7 @@
static bool hasChildrenInFlags(const uint8_t flags);
static int getAttributeAddressAndForwardPointer(const uint8_t* const dict, const uint8_t flags,
int *pos);
- static int getTerminalPosition(const uint8_t* const root, const uint16_t* const inWord,
+ static int getTerminalPosition(const uint8_t* const root, const int32_t* const inWord,
const int length);
static int getWordAtAddress(const uint8_t* const root, const int address, const int maxDepth,
uint16_t* outWord);
@@ -304,7 +304,7 @@
// This function gets the byte position of the last chargroup of the exact matching word in the
// dictionary. If no match is found, it returns NOT_VALID_WORD.
inline int BinaryFormat::getTerminalPosition(const uint8_t* const root,
- const uint16_t* const inWord, const int length) {
+ const int32_t* const inWord, const int length) {
int pos = 0;
int wordPos = 0;
@@ -313,7 +313,7 @@
// there was no match (or we would have found it).
if (wordPos > length) return NOT_VALID_WORD;
int charGroupCount = BinaryFormat::getGroupCountAndForwardPointer(root, &pos);
- const uint16_t wChar = inWord[wordPos];
+ const int32_t wChar = inWord[wordPos];
while (true) {
// If there are no more character groups in this node, it means we could not
// find a matching character for this depth, therefore there is no match.
diff --git a/native/jni/src/dictionary.cpp b/native/jni/src/dictionary.cpp
index 90ec207..9dc2072 100644
--- a/native/jni/src/dictionary.cpp
+++ b/native/jni/src/dictionary.cpp
@@ -54,7 +54,7 @@
delete mBigramDictionary;
}
-bool Dictionary::isValidWord(unsigned short *word, int length) {
+bool Dictionary::isValidWord(const int32_t *word, int length) {
return mUnigramDictionary->isValidWord(word, length);
}
diff --git a/native/jni/src/dictionary.h b/native/jni/src/dictionary.h
index 66a5c21..dea5763 100644
--- a/native/jni/src/dictionary.h
+++ b/native/jni/src/dictionary.h
@@ -40,13 +40,13 @@
codesSize, useFullEditDistance, outWords, frequencies);
}
- int getBigrams(unsigned short *word, int length, int *codes, int codesSize,
+ int getBigrams(const int32_t *word, int length, int *codes, int codesSize,
unsigned short *outWords, int *frequencies, int maxWordLength, int maxBigrams) {
return mBigramDictionary->getBigrams(word, length, codes, codesSize, outWords, frequencies,
maxWordLength, maxBigrams);
}
- bool isValidWord(unsigned short *word, int length);
+ bool isValidWord(const int32_t *word, int length);
void *getDict() { return (void *)mDict; }
int getDictSize() { return mDictSize; }
int getMmapFd() { return mMmapFd; }
diff --git a/native/jni/src/unigram_dictionary.cpp b/native/jni/src/unigram_dictionary.cpp
index ab8570e..05c124b 100644
--- a/native/jni/src/unigram_dictionary.cpp
+++ b/native/jni/src/unigram_dictionary.cpp
@@ -730,7 +730,7 @@
return maxFreq;
}
-bool UnigramDictionary::isValidWord(const uint16_t* const inWord, const int length) const {
+bool UnigramDictionary::isValidWord(const int32_t* const inWord, const int length) const {
return NOT_VALID_WORD != BinaryFormat::getTerminalPosition(DICT_ROOT, inWord, length);
}
diff --git a/native/jni/src/unigram_dictionary.h b/native/jni/src/unigram_dictionary.h
index 4479cd9..b09c000 100644
--- a/native/jni/src/unigram_dictionary.h
+++ b/native/jni/src/unigram_dictionary.h
@@ -71,7 +71,7 @@
UnigramDictionary(const uint8_t* const streamStart, int typedLetterMultipler,
int fullWordMultiplier, int maxWordLength, int maxWords, const unsigned int flags);
- bool isValidWord(const uint16_t* const inWord, const int length) const;
+ bool isValidWord(const int32_t* const inWord, const int length) const;
int getBigramPosition(int pos, unsigned short *word, int offset, int length) const;
int getSuggestions(ProximityInfo *proximityInfo, WordsPriorityQueuePool *queuePool,
Correction *correction, const int *xcoordinates,