Home | History | Annotate | Download | only in dictionary 
      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 "dictionary/interface/dictionary_header_structure_policy.h"
     25 #include "dictionary/interface/dictionary_structure_with_buffer_policy.h"
     26 #include "dictionary/interface/ngram_listener.h"
     27 #include "dictionary/property/historical_info.h"
     28 #include "dictionary/property/word_property.h"
     29 #include "suggest/core/suggest_interface.h"
     30 #include "utils/int_array_view.h"
     31 
     32 namespace latinime {
     33 
     34 class DictionaryStructureWithBufferPolicy;
     35 class DicTraverseSession;
     36 class NgramContext;
     37 class ProximityInfo;
     38 class SuggestionResults;
     39 class SuggestOptions;
     40 
     41 class Dictionary {
     42  public:
     43     // Taken from SuggestedWords.java
     44     static const int KIND_MASK_KIND = 0xFF; // Mask to get only the kind
     45     static const int KIND_TYPED = 0; // What user typed
     46     static const int KIND_CORRECTION = 1; // Simple correction/suggestion
     47     static const int KIND_COMPLETION = 2; // Completion (suggestion with appended chars)
     48     static const int KIND_WHITELIST = 3; // Whitelisted word
     49     static const int KIND_BLACKLIST = 4; // Blacklisted word
     50     static const int KIND_HARDCODED = 5; // Hardcoded suggestion, e.g. punctuation
     51     static const int KIND_APP_DEFINED = 6; // Suggested by the application
     52     static const int KIND_SHORTCUT = 7; // A shortcut
     53     static const int KIND_PREDICTION = 8; // A prediction (== a suggestion with no input)
     54     // KIND_RESUMED: A resumed suggestion (comes from a span, currently this type is used only
     55     // in java for re-correction)
     56     static const int KIND_RESUMED = 9;
     57     static const int KIND_OOV_CORRECTION = 10; // Most probable string correction
     58 
     59     static const int KIND_MASK_FLAGS = 0xFFFFFF00; // Mask to get the flags
     60     static const int KIND_FLAG_POSSIBLY_OFFENSIVE = 0x80000000;
     61     static const int KIND_FLAG_EXACT_MATCH = 0x40000000;
     62     static const int KIND_FLAG_EXACT_MATCH_WITH_INTENTIONAL_OMISSION = 0x20000000;
     63     static const int KIND_FLAG_APPROPRIATE_FOR_AUTOCORRECTION = 0x10000000;
     64 
     65     Dictionary(JNIEnv *env, DictionaryStructureWithBufferPolicy::StructurePolicyPtr
     66             dictionaryStructureWithBufferPolicy);
     67 
     68     void getSuggestions(ProximityInfo *proximityInfo, DicTraverseSession *traverseSession,
     69             int *xcoordinates, int *ycoordinates, int *times, int *pointerIds, int *inputCodePoints,
     70             int inputSize, const NgramContext *const ngramContext,
     71             const SuggestOptions *const suggestOptions, const float weightOfLangModelVsSpatialModel,
     72             SuggestionResults *const outSuggestionResults) const;
     73 
     74     void getPredictions(const NgramContext *const ngramContext,
     75             SuggestionResults *const outSuggestionResults) const;
     76 
     77     int getProbability(const CodePointArrayView codePoints) const;
     78 
     79     int getMaxProbabilityOfExactMatches(const CodePointArrayView codePoints) const;
     80 
     81     int getNgramProbability(const NgramContext *const ngramContext,
     82             const CodePointArrayView codePoints) const;
     83 
     84     bool addUnigramEntry(const CodePointArrayView codePoints,
     85             const UnigramProperty *const unigramProperty);
     86 
     87     bool removeUnigramEntry(const CodePointArrayView codePoints);
     88 
     89     bool addNgramEntry(const NgramProperty *const ngramProperty);
     90 
     91     bool removeNgramEntry(const NgramContext *const ngramContext,
     92             const CodePointArrayView codePoints);
     93 
     94     bool updateEntriesForWordWithNgramContext(const NgramContext *const ngramContext,
     95             const CodePointArrayView codePoints, const bool isValidWord,
     96             const HistoricalInfo historicalInfo);
     97 
     98     bool flush(const char *const filePath);
     99 
    100     bool flushWithGC(const char *const filePath);
    101 
    102     bool needsToRunGC(const bool mindsBlockByGC);
    103 
    104     void getProperty(const char *const query, const int queryLength, char *const outResult,
    105             const int maxResultLength);
    106 
    107     const WordProperty getWordProperty(const CodePointArrayView codePoints);
    108 
    109     // Method to iterate all words in the dictionary.
    110     // The returned token has to be used to get the next word. If token is 0, this method newly
    111     // starts iterating the dictionary.
    112     int getNextWordAndNextToken(const int token, int *const outCodePoints,
    113             int *const outCodePointCount);
    114 
    115     const DictionaryStructureWithBufferPolicy *getDictionaryStructurePolicy() const {
    116         return mDictionaryStructureWithBufferPolicy.get();
    117     }
    118 
    119  private:
    120     DISALLOW_IMPLICIT_CONSTRUCTORS(Dictionary);
    121 
    122     typedef std::unique_ptr<SuggestInterface> SuggestInterfacePtr;
    123 
    124     class NgramListenerForPrediction : public NgramListener {
    125      public:
    126         NgramListenerForPrediction(const NgramContext *const ngramContext,
    127                 const WordIdArrayView prevWordIds, SuggestionResults *const suggestionResults,
    128                 const DictionaryStructureWithBufferPolicy *const dictStructurePolicy);
    129         virtual void onVisitEntry(const int ngramProbability, const int targetWordId);
    130 
    131      private:
    132         DISALLOW_IMPLICIT_CONSTRUCTORS(NgramListenerForPrediction);
    133 
    134         const NgramContext *const mNgramContext;
    135         const WordIdArrayView mPrevWordIds;
    136         SuggestionResults *const mSuggestionResults;
    137         const DictionaryStructureWithBufferPolicy *const mDictStructurePolicy;
    138     };
    139 
    140     static const int HEADER_ATTRIBUTE_BUFFER_SIZE;
    141 
    142     const DictionaryStructureWithBufferPolicy::StructurePolicyPtr
    143             mDictionaryStructureWithBufferPolicy;
    144     const SuggestInterfacePtr mGestureSuggest;
    145     const SuggestInterfacePtr mTypingSuggest;
    146 
    147     void logDictionaryInfo(JNIEnv *const env) const;
    148 };
    149 } // namespace latinime
    150 #endif // LATINIME_DICTIONARY_H
    151