blob: f3bdb310d8f7ccb2188dea64a386935294c3a679 [file] [log] [blame]
The Android Open Source Project923bf412009-03-13 15:11:42 -07001/*
Ken Wakasa0bbb9172012-07-25 17:51:43 +09002 * Copyright (C) 2009, The Android Open Source Project
3 *
4 * Licensed under the Apache License, Version 2.0 (the "License");
5 * you may not use this file except in compliance with the License.
6 * You may obtain a copy of the License at
7 *
8 * http://www.apache.org/licenses/LICENSE-2.0
9 *
10 * Unless required by applicable law or agreed to in writing, software
11 * distributed under the License is distributed on an "AS IS" BASIS,
12 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
13 * See the License for the specific language governing permissions and
14 * limitations under the License.
15 */
The Android Open Source Project923bf412009-03-13 15:11:42 -070016
satoke808e432010-12-02 14:53:24 +090017#define LOG_TAG "LatinIME: dictionary.cpp"
18
Ken Wakasa77e8e812012-08-02 19:48:08 +090019#include <stdint.h>
20
21#include "bigram_dictionary.h"
Jean Chalard46a1eec2012-02-27 19:48:47 +090022#include "binary_format.h"
Ken Wakasa3b088a22012-05-16 23:05:32 +090023#include "defines.h"
The Android Open Source Project923bf412009-03-13 15:11:42 -070024#include "dictionary.h"
Satoshi Kataokae9f3e182012-08-09 23:23:08 +090025#include "dic_traverse_wrapper.h"
Satoshi Kataokadeb09872012-07-03 17:45:50 +090026#include "gesture_decoder_wrapper.h"
Ken Wakasa77e8e812012-08-02 19:48:08 +090027#include "unigram_dictionary.h"
satokd4952c82010-12-01 19:09:29 +090028
The Android Open Source Project923bf412009-03-13 15:11:42 -070029namespace latinime {
30
satok8fbd5522011-02-22 17:28:55 +090031// TODO: Change the type of all keyCodes to uint32_t
Ken Wakasae90b3332011-01-07 15:01:51 +090032Dictionary::Dictionary(void *dict, int dictSize, int mmapFd, int dictBufAdjust,
33 int typedLetterMultiplier, int fullWordMultiplier,
Jean Chalardb7d7c5a2012-07-11 11:31:48 +090034 int maxWordLength, int maxWords, int maxPredictions)
Satoshi Kataokae9f3e182012-08-09 23:23:08 +090035 : mDict((unsigned char*) dict),
36 mOffsetDict(((unsigned char*) dict) + BinaryFormat::getHeaderSize(mDict)),
37 mDictSize(dictSize), mMmapFd(mmapFd), mDictBufAdjust(dictBufAdjust) {
satok662fe692010-12-08 17:05:39 +090038 if (DEBUG_DICT) {
39 if (MAX_WORD_LENGTH_INTERNAL < maxWordLength) {
satok9fb6f472012-01-13 18:01:22 +090040 AKLOGI("Max word length (%d) is greater than %d",
satok662fe692010-12-08 17:05:39 +090041 maxWordLength, MAX_WORD_LENGTH_INTERNAL);
satok9fb6f472012-01-13 18:01:22 +090042 AKLOGI("IN NATIVE SUGGEST Version: %d", (mDict[0] & 0xFF));
satok662fe692010-12-08 17:05:39 +090043 }
satok715514d2010-12-02 20:19:59 +090044 }
Jean Chalardcd274b12012-04-06 18:26:00 +090045 const unsigned int options = BinaryFormat::getFlags(mDict);
Satoshi Kataokae9f3e182012-08-09 23:23:08 +090046 mUnigramDictionary = new UnigramDictionary(mOffsetDict, typedLetterMultiplier,
Jean Chalardcd274b12012-04-06 18:26:00 +090047 fullWordMultiplier, maxWordLength, maxWords, options);
Satoshi Kataokae9f3e182012-08-09 23:23:08 +090048 mBigramDictionary = new BigramDictionary(mOffsetDict, maxWordLength, maxPredictions);
Satoshi Kataokadeb09872012-07-03 17:45:50 +090049 mGestureDecoder = new GestureDecoderWrapper(maxWordLength, maxWords);
Satoshi Kataokaefb63242012-06-27 14:52:40 +090050 mGestureDecoder->setDict(mUnigramDictionary, mBigramDictionary,
Satoshi Kataokae9f3e182012-08-09 23:23:08 +090051 mOffsetDict /* dict root */, 0 /* root pos */);
The Android Open Source Project923bf412009-03-13 15:11:42 -070052}
53
satok662fe692010-12-08 17:05:39 +090054Dictionary::~Dictionary() {
satok30088252010-12-01 21:22:15 +090055 delete mUnigramDictionary;
56 delete mBigramDictionary;
Ken Wakasa8658e552012-06-30 08:53:33 +090057 delete mGestureDecoder;
The Android Open Source Project923bf412009-03-13 15:11:42 -070058}
satoke808e432010-12-02 14:53:24 +090059
Satoshi Kataoka91278112012-08-08 21:23:25 +090060int Dictionary::getSuggestions(ProximityInfo *proximityInfo, void *traverseSession,
61 int *xcoordinates, int *ycoordinates, int *times, int *pointerIds,
62 int *codes, int codesSize, int *prevWordChars,
Ken Wakasa77e8e812012-08-02 19:48:08 +090063 int prevWordLength, int commitPoint, bool isGesture,
64 bool useFullEditDistance, unsigned short *outWords,
65 int *frequencies, int *spaceIndices, int *outputTypes) {
66 int result = 0;
67 if (isGesture) {
Satoshi Kataokae9f3e182012-08-09 23:23:08 +090068 DicTraverseWrapper::initDicTraverseSession(
69 traverseSession, this, prevWordChars, prevWordLength);
Satoshi Kataoka91278112012-08-08 21:23:25 +090070 result = mGestureDecoder->getSuggestions(proximityInfo, traverseSession,
71 xcoordinates, ycoordinates, times, pointerIds, codes, codesSize, commitPoint,
Ken Wakasa77e8e812012-08-02 19:48:08 +090072 outWords, frequencies, spaceIndices, outputTypes);
Satoshi Kataoka586b0ca2012-08-06 11:20:54 +090073 if (DEBUG_DICT) {
74 DUMP_RESULT(outWords, frequencies, 18 /* MAX_WORDS */, MAX_WORD_LENGTH_INTERNAL);
75 }
Ken Wakasa77e8e812012-08-02 19:48:08 +090076 return result;
77 } else {
78 std::map<int, int> bigramMap;
79 uint8_t bigramFilter[BIGRAM_FILTER_BYTE_SIZE];
80 mBigramDictionary->fillBigramAddressToFrequencyMapAndFilter(prevWordChars,
81 prevWordLength, &bigramMap, bigramFilter);
82 result = mUnigramDictionary->getSuggestions(proximityInfo, xcoordinates,
83 ycoordinates, codes, codesSize, &bigramMap, bigramFilter,
84 useFullEditDistance, outWords, frequencies, outputTypes);
85 return result;
86 }
87}
88
89int Dictionary::getBigrams(const int32_t *word, int length, int *codes, int codesSize,
90 unsigned short *outWords, int *frequencies, int *outputTypes) const {
91 if (length <= 0) return 0;
92 return mBigramDictionary->getBigrams(word, length, codes, codesSize, outWords, frequencies,
93 outputTypes);
94}
95
satokb1ed1d42012-06-14 16:35:23 -070096int Dictionary::getFrequency(const int32_t *word, int length) const {
Satoshi Kataoka2f854e12012-05-29 15:58:13 +090097 return mUnigramDictionary->getFrequency(word, length);
satoke808e432010-12-02 14:53:24 +090098}
99
Tom Ouyang4d289d32012-04-26 23:50:21 -0700100bool Dictionary::isValidBigram(const int32_t *word1, int length1, const int32_t *word2,
satokb1ed1d42012-06-14 16:35:23 -0700101 int length2) const {
Tom Ouyang4d289d32012-04-26 23:50:21 -0700102 return mBigramDictionary->isValidBigram(word1, length1, word2, length2);
103}
The Android Open Source Project923bf412009-03-13 15:11:42 -0700104} // namespace latinime