Merge "New dict format, step 1" (3f438551) · Commits · e / os / android_packages_inputmethods_LatinIME

native/src/unigram_dictionary.cpp

+13 −13

Original line number	Diff line number	Diff line
		@@ -16,8 +16,6 @@
		*/

		#include <assert.h>
		#include <fcntl.h>
		#include <stdio.h>
		#include <string.h>

		#define LOG_TAG "LatinIME: unigram_dictionary.cpp"
		@@ -34,10 +32,12 @@ const UnigramDictionary::digraph_t UnigramDictionary::GERMAN_UMLAUT_DIGRAPHS[] =
		{ 'o', 'e' },
		{ 'u', 'e' } };

		UnigramDictionary::UnigramDictionary(const unsigned char *dict, int typedLetterMultiplier,
		// TODO: check the header
		UnigramDictionary::UnigramDictionary(const uint8_t* const streamStart, int typedLetterMultiplier,
		int fullWordMultiplier, int maxWordLength, int maxWords, int maxProximityChars,
		const bool isLatestDictVersion)
		: DICT(dict), MAX_WORD_LENGTH(maxWordLength), MAX_WORDS(maxWords),
		: DICT_ROOT(streamStart),
		MAX_WORD_LENGTH(maxWordLength), MAX_WORDS(maxWords),
		MAX_PROXIMITY_CHARS(maxProximityChars), IS_LATEST_DICT_VERSION(isLatestDictVersion),
		TYPED_LETTER_MULTIPLIER(typedLetterMultiplier), FULL_WORD_MULTIPLIER(fullWordMultiplier),
		ROOT_POS(isLatestDictVersion ? DICTIONARY_HEADER_SIZE : 0),
		@@ -363,7 +363,7 @@ void UnigramDictionary::getSuggestionCandidates(const int skipPos,
		}
		int rootPosition = ROOT_POS;
		// Get the number of child of root, then increment the position
		int childCount = Dictionary::getCount(DICT, &rootPosition);
		int childCount = Dictionary::getCount(DICT_ROOT, &rootPosition);
		int depth = 0;

		mStackChildCount[0] = childCount;
		@@ -562,7 +562,7 @@ void UnigramDictionary::getWordsOld(const int initialPos, const int inputLength,
		const int excessivePos, const int transposedPos,int *nextLetters,
		const int nextLettersSize) {
		int initialPosition = initialPos;
		const int count = Dictionary::getCount(DICT, &initialPosition);
		const int count = Dictionary::getCount(DICT_ROOT, &initialPosition);
		getWordsRec(count, initialPosition, 0,
		min(inputLength * MAX_DEPTH_MULTIPLIER, MAX_WORD_LENGTH),
		mInputLength <= 0, 1, 0, 0, skipPos, excessivePos, transposedPos, nextLetters,
		@@ -770,8 +770,8 @@ inline bool UnigramDictionary::processCurrentNode(const int pos, const int depth

		if (excessivePos == depth && inputIndex < mInputLength - 1) ++inputIndex;

		*nextSiblingPosition = Dictionary::setDictionaryValues(DICT, IS_LATEST_DICT_VERSION, pos, &c,
		&childPosition, &terminal, &freq);
		*nextSiblingPosition = Dictionary::setDictionaryValues(DICT_ROOT, IS_LATEST_DICT_VERSION, pos,
		&c, &childPosition, &terminal, &freq);

		const bool needsToTraverseChildrenNodes = childPosition != 0;

		@@ -829,7 +829,7 @@ inline bool UnigramDictionary::processCurrentNode(const int pos, const int depth
		*newTraverseAllNodes = true;
		}
		// get the count of nodes and increment childAddress.
		*newCount = Dictionary::getCount(DICT, &childPosition);
		*newCount = Dictionary::getCount(DICT_ROOT, &childPosition);
		*newChildPosition = childPosition;
		if (DEBUG_DICT) assert(needsToTraverseChildrenNodes);
		return needsToTraverseChildrenNodes;
		@@ -838,7 +838,7 @@ inline bool UnigramDictionary::processCurrentNode(const int pos, const int depth
		inline int UnigramDictionary::getBestWordFreq(const int startInputIndex, const int inputLength,
		unsigned short *word) {
		int pos = ROOT_POS;
		int count = Dictionary::getCount(DICT, &pos);
		int count = Dictionary::getCount(DICT_ROOT, &pos);
		int maxFreq = 0;
		int depth = 0;
		unsigned short newWord[MAX_WORD_LENGTH_INTERNAL];
		@@ -894,8 +894,8 @@ inline bool UnigramDictionary::processCurrentNodeForExactMatch(const int firstCh
		const int inputIndex = startInputIndex + depth;
		const int *currentChars = getInputCharsAt(inputIndex);
		unsigned short c;
		*siblingPos = Dictionary::setDictionaryValues(DICT, IS_LATEST_DICT_VERSION, firstChildPos, &c,
		newChildPosition, newTerminal, newFreq);
		*siblingPos = Dictionary::setDictionaryValues(DICT_ROOT, IS_LATEST_DICT_VERSION, firstChildPos,
		&c, newChildPosition, newTerminal, newFreq);
		const unsigned int inputC = currentChars[0];
		if (DEBUG_DICT) {
		assert(inputC <= U_SHORT_MAX);
		@@ -912,7 +912,7 @@ inline bool UnigramDictionary::processCurrentNodeForExactMatch(const int firstCh
		}
		}
		if (hasChild) {
		*newCount = Dictionary::getCount(DICT, newChildPosition);
		*newCount = Dictionary::getCount(DICT_ROOT, newChildPosition);
		return true;
		} else {
		return false;

native/src/unigram_dictionary.h

+10 −3

Original line number	Diff line number	Diff line
		@@ -17,9 +17,14 @@
		#ifndef LATINIME_UNIGRAM_DICTIONARY_H
		#define LATINIME_UNIGRAM_DICTIONARY_H

		#include <stdint.h>
		#include "defines.h"
		#include "proximity_info.h"

		#ifndef NULL
		#define NULL 0
		#endif

		namespace latinime {

		class UnigramDictionary {
		@@ -31,8 +36,9 @@ class UnigramDictionary {
		} ProximityType;

		public:
		UnigramDictionary(const unsigned char *dict, int typedLetterMultipler, int fullWordMultiplier,
		int maxWordLength, int maxWords, int maxProximityChars, const bool isLatestDictVersion);
		UnigramDictionary(const uint8_t* const streamStart, int typedLetterMultipler,
		int fullWordMultiplier, int maxWordLength, int maxWords, int maxProximityChars,
		const bool isLatestDictVersion);
		int getSuggestions(const ProximityInfo proximityInfo, const int xcoordinates,
		const int ycoordinates, const int codes, const int codesSize, const int flags,
		unsigned short outWords, int frequencies);
		@@ -104,7 +110,8 @@ private:
		inline const int* getInputCharsAt(const int index) const {
		return mInputCodes + (index * MAX_PROXIMITY_CHARS);
		}
		const unsigned char *DICT;

		const uint8_t* const DICT_ROOT;
		const int MAX_WORD_LENGTH;
		const int MAX_WORDS;
		const int MAX_PROXIMITY_CHARS;