1/*
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
20#include <memory>
21
22#include "defines.h"
23#include "jni.h"
24#include "suggest/core/dictionary/ngram_listener.h"
25#include "suggest/core/dictionary/property/word_property.h"
26#include "suggest/core/policy/dictionary_header_structure_policy.h"
27#include "suggest/core/policy/dictionary_structure_with_buffer_policy.h"
28#include "suggest/core/suggest_interface.h"
29
30namespace latinime {
31
32class DictionaryStructureWithBufferPolicy;
33class DicTraverseSession;
34class PrevWordsInfo;
35class ProximityInfo;
36class SuggestionResults;
37class SuggestOptions;
38
39class Dictionary {
40 public:
41    // Taken from SuggestedWords.java
42    static const int KIND_MASK_KIND = 0xFF; // Mask to get only the kind
43    static const int KIND_TYPED = 0; // What user typed
44    static const int KIND_CORRECTION = 1; // Simple correction/suggestion
45    static const int KIND_COMPLETION = 2; // Completion (suggestion with appended chars)
46    static const int KIND_WHITELIST = 3; // Whitelisted word
47    static const int KIND_BLACKLIST = 4; // Blacklisted word
48    static const int KIND_HARDCODED = 5; // Hardcoded suggestion, e.g. punctuation
49    static const int KIND_APP_DEFINED = 6; // Suggested by the application
50    static const int KIND_SHORTCUT = 7; // A shortcut
51    static const int KIND_PREDICTION = 8; // A prediction (== a suggestion with no input)
52    // KIND_RESUMED: A resumed suggestion (comes from a span, currently this type is used only
53    // in java for re-correction)
54    static const int KIND_RESUMED = 9;
55    static const int KIND_OOV_CORRECTION = 10; // Most probable string correction
56
57    static const int KIND_MASK_FLAGS = 0xFFFFFF00; // Mask to get the flags
58    static const int KIND_FLAG_POSSIBLY_OFFENSIVE = 0x80000000;
59    static const int KIND_FLAG_EXACT_MATCH = 0x40000000;
60    static const int KIND_FLAG_EXACT_MATCH_WITH_INTENTIONAL_OMISSION = 0x20000000;
61
62    Dictionary(JNIEnv *env, DictionaryStructureWithBufferPolicy::StructurePolicyPtr
63            dictionaryStructureWithBufferPolicy);
64
65    void getSuggestions(ProximityInfo *proximityInfo, DicTraverseSession *traverseSession,
66            int *xcoordinates, int *ycoordinates, int *times, int *pointerIds, int *inputCodePoints,
67            int inputSize, const PrevWordsInfo *const prevWordsInfo,
68            const SuggestOptions *const suggestOptions, const float languageWeight,
69            SuggestionResults *const outSuggestionResults) const;
70
71    void getPredictions(const PrevWordsInfo *const prevWordsInfo,
72            SuggestionResults *const outSuggestionResults) const;
73
74    int getProbability(const int *word, int length) const;
75
76    int getMaxProbabilityOfExactMatches(const int *word, int length) const;
77
78    int getNgramProbability(const PrevWordsInfo *const prevWordsInfo,
79            const int *word, int length) const;
80
81    bool addUnigramEntry(const int *const codePoints, const int codePointCount,
82            const UnigramProperty *const unigramProperty);
83
84    bool removeUnigramEntry(const int *const codePoints, const int codePointCount);
85
86    bool addNgramEntry(const PrevWordsInfo *const prevWordsInfo,
87            const BigramProperty *const bigramProperty);
88
89    bool removeNgramEntry(const PrevWordsInfo *const prevWordsInfo, const int *const word,
90            const int length);
91
92    bool flush(const char *const filePath);
93
94    bool flushWithGC(const char *const filePath);
95
96    bool needsToRunGC(const bool mindsBlockByGC);
97
98    void getProperty(const char *const query, const int queryLength, char *const outResult,
99            const int maxResultLength);
100
101    const WordProperty getWordProperty(const int *const codePoints, const int codePointCount);
102
103    // Method to iterate all words in the dictionary.
104    // The returned token has to be used to get the next word. If token is 0, this method newly
105    // starts iterating the dictionary.
106    int getNextWordAndNextToken(const int token, int *const outCodePoints,
107            int *const outCodePointCount);
108
109    const DictionaryStructureWithBufferPolicy *getDictionaryStructurePolicy() const {
110        return mDictionaryStructureWithBufferPolicy.get();
111    }
112
113 private:
114    DISALLOW_IMPLICIT_CONSTRUCTORS(Dictionary);
115
116    typedef std::unique_ptr<SuggestInterface> SuggestInterfacePtr;
117
118    class NgramListenerForPrediction : public NgramListener {
119     public:
120        NgramListenerForPrediction(const PrevWordsInfo *const prevWordsInfo,
121                SuggestionResults *const suggestionResults,
122                const DictionaryStructureWithBufferPolicy *const dictStructurePolicy);
123        virtual void onVisitEntry(const int ngramProbability, const int targetPtNodePos);
124
125     private:
126        DISALLOW_IMPLICIT_CONSTRUCTORS(NgramListenerForPrediction);
127
128        const PrevWordsInfo *const mPrevWordsInfo;
129        SuggestionResults *const mSuggestionResults;
130        const DictionaryStructureWithBufferPolicy *const mDictStructurePolicy;
131    };
132
133    static const int HEADER_ATTRIBUTE_BUFFER_SIZE;
134
135    const DictionaryStructureWithBufferPolicy::StructurePolicyPtr
136            mDictionaryStructureWithBufferPolicy;
137    const SuggestInterfacePtr mGestureSuggest;
138    const SuggestInterfacePtr mTypingSuggest;
139
140    void logDictionaryInfo(JNIEnv *const env) const;
141};
142} // namespace latinime
143#endif // LATINIME_DICTIONARY_H
144