blob: c8a21e0ae0b6f477d04af3365f98aae47a76f1ac [file] [log] [blame]
The Android Open Source Project923bf412009-03-13 15:11:42 -07001/*
2 * 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 */
16
17#ifndef LATINIME_DICTIONARY_H
18#define LATINIME_DICTIONARY_H
19
Jean Chalard1ff8dc42012-05-02 16:00:24 +090020#include <map>
21
satok30088252010-12-01 21:22:15 +090022#include "bigram_dictionary.h"
satokd24df432011-07-14 15:43:42 +090023#include "char_utils.h"
satoke808e432010-12-02 14:53:24 +090024#include "defines.h"
Ken Wakasa8658e552012-06-30 08:53:33 +090025#include "incremental_decoder_interface.h"
satok8fbd5522011-02-22 17:28:55 +090026#include "proximity_info.h"
satok30088252010-12-01 21:22:15 +090027#include "unigram_dictionary.h"
satoka7e5a5a2011-12-15 16:49:12 +090028#include "words_priority_queue_pool.h"
satok30088252010-12-01 21:22:15 +090029
The Android Open Source Project923bf412009-03-13 15:11:42 -070030namespace latinime {
31
32class Dictionary {
Ken Wakasae12e9b52012-01-06 12:24:38 +090033 public:
Ken Wakasae90b3332011-01-07 15:01:51 +090034 Dictionary(void *dict, int dictSize, int mmapFd, int dictBufAdjust, int typedLetterMultipler,
Jean Chalardb7d7c5a2012-07-11 11:31:48 +090035 int fullWordMultiplier, int maxWordLength, int maxWords, int maxPredictions);
satok1147c7b2011-12-14 15:04:58 +090036
satok8fbd5522011-02-22 17:28:55 +090037 int getSuggestions(ProximityInfo *proximityInfo, int *xcoordinates, int *ycoordinates,
Satoshi Kataoka73680092012-06-25 17:44:54 +090038 int *times, int *pointerIds, int *codes, int codesSize, int *prevWordChars,
Jean Chalard05efe572012-06-27 17:31:09 +090039 int prevWordLength, int commitPoint, bool isGesture,
Satoshi Kataoka73680092012-06-25 17:44:54 +090040 bool useFullEditDistance, unsigned short *outWords,
41 int *frequencies, int *spaceIndices) {
42 int result = 0;
Satoshi Kataoka91eb4d82012-06-26 16:39:07 +090043 if (isGesture) {
44 mGestureDecoder->setPrevWord(prevWordChars, prevWordLength);
45 result = mGestureDecoder->getSuggestions(proximityInfo, xcoordinates, ycoordinates,
Jean Chalard05efe572012-06-27 17:31:09 +090046 times, pointerIds, codes, codesSize, commitPoint,
Satoshi Kataoka91eb4d82012-06-26 16:39:07 +090047 outWords, frequencies, spaceIndices);
Satoshi Kataokadeb09872012-07-03 17:45:50 +090048 return result;
Satoshi Kataoka91eb4d82012-06-26 16:39:07 +090049 } else {
50 std::map<int, int> bigramMap;
51 uint8_t bigramFilter[BIGRAM_FILTER_BYTE_SIZE];
52 mBigramDictionary->fillBigramAddressToFrequencyMapAndFilter(prevWordChars,
53 prevWordLength, &bigramMap, bigramFilter);
54 result = mUnigramDictionary->getSuggestions(proximityInfo, xcoordinates,
55 ycoordinates, codes, codesSize, &bigramMap, bigramFilter,
56 useFullEditDistance, outWords, frequencies);
Satoshi Kataokadeb09872012-07-03 17:45:50 +090057 return result;
Satoshi Kataoka91eb4d82012-06-26 16:39:07 +090058 }
satok30088252010-12-01 21:22:15 +090059 }
60
Jean Chalard522a04e2012-04-23 15:37:07 +090061 int getBigrams(const int32_t *word, int length, int *codes, int codesSize,
Jean Chalard6a5d17c2012-07-11 11:55:25 +090062 unsigned short *outWords, int *frequencies) const {
Jean Chalard3b576312012-07-11 16:47:22 +090063 if (length <= 0) return 0;
Jean Chalard6a5d17c2012-07-11 11:55:25 +090064 return mBigramDictionary->getBigrams(word, length, codes, codesSize, outWords, frequencies);
satok30088252010-12-01 21:22:15 +090065 }
satok8fbd5522011-02-22 17:28:55 +090066
satokb1ed1d42012-06-14 16:35:23 -070067 int getFrequency(const int32_t *word, int length) const;
68 bool isValidBigram(const int32_t *word1, int length1, const int32_t *word2, int length2) const;
69 void *getDict() const { return (void *)mDict; }
70 int getDictSize() const { return mDictSize; }
71 int getMmapFd() const { return mMmapFd; }
72 int getDictBufAdjust() const { return mDictBufAdjust; }
The Android Open Source Project923bf412009-03-13 15:11:42 -070073 ~Dictionary();
Amith Yamasanicc3e5c72009-03-31 10:51:17 -070074
satoke808e432010-12-02 14:53:24 +090075 // public static utility methods
76 // static inline methods should be defined in the header file
satok18c28f42010-12-02 18:11:54 +090077 static int wideStrLen(unsigned short *str);
Jean Chalard581335c2011-06-17 12:45:17 +090078
Ken Wakasae12e9b52012-01-06 12:24:38 +090079 private:
satok1bc038c2012-06-14 11:25:50 -070080 DISALLOW_IMPLICIT_CONSTRUCTORS(Dictionary);
Ken Wakasae90b3332011-01-07 15:01:51 +090081 const unsigned char *mDict;
82
83 // Used only for the mmap version of dictionary loading, but we use these as dummy variables
84 // also for the malloc version.
85 const int mDictSize;
86 const int mMmapFd;
87 const int mDictBufAdjust;
88
satok1bc038c2012-06-14 11:25:50 -070089 const UnigramDictionary *mUnigramDictionary;
satokb1ed1d42012-06-14 16:35:23 -070090 const BigramDictionary *mBigramDictionary;
Ken Wakasa8658e552012-06-30 08:53:33 +090091 IncrementalDecoderInterface *mGestureDecoder;
The Android Open Source Project923bf412009-03-13 15:11:42 -070092};
93
satoke808e432010-12-02 14:53:24 +090094// public static utility methods
95// static inline methods should be defined in the header file
satok18c28f42010-12-02 18:11:54 +090096inline int Dictionary::wideStrLen(unsigned short *str) {
97 if (!str) return 0;
98 unsigned short *end = str;
99 while (*end)
100 end++;
101 return end - str;
102}
Ken Wakasace9e52a2011-06-18 13:09:55 +0900103} // namespace latinime
104
The Android Open Source Project923bf412009-03-13 15:11:42 -0700105#endif // LATINIME_DICTIONARY_H