1 /*
2  * Copyright (C) 2013, 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_DYNAMIC_PT_UPDATING_HELPER_H
18 #define LATINIME_DYNAMIC_PT_UPDATING_HELPER_H
19 
20 #include "defines.h"
21 #include "dictionary/structure/pt_common/pt_node_params.h"
22 #include "utils/int_array_view.h"
23 
24 namespace latinime {
25 
26 class NgramProperty;
27 class BufferWithExtendableBuffer;
28 class DynamicPtReadingHelper;
29 class PtNodeReader;
30 class PtNodeWriter;
31 class UnigramProperty;
32 
33 class DynamicPtUpdatingHelper {
34  public:
DynamicPtUpdatingHelper(BufferWithExtendableBuffer * const buffer,const PtNodeReader * const ptNodeReader,PtNodeWriter * const ptNodeWriter)35     DynamicPtUpdatingHelper(BufferWithExtendableBuffer *const buffer,
36             const PtNodeReader *const ptNodeReader, PtNodeWriter *const ptNodeWriter)
37             : mBuffer(buffer), mPtNodeReader(ptNodeReader), mPtNodeWriter(ptNodeWriter) {}
38 
~DynamicPtUpdatingHelper()39     ~DynamicPtUpdatingHelper() {}
40 
41     // Add a word to the dictionary. If the word already exists, update the probability.
42     bool addUnigramWord(DynamicPtReadingHelper *const readingHelper,
43             const CodePointArrayView wordCodePoints, const UnigramProperty *const unigramProperty,
44             bool *const outAddedNewUnigram);
45 
46     // TODO: Remove after stopping supporting v402.
47     // Add an n-gram entry.
48     bool addNgramEntry(const PtNodePosArrayView prevWordsPtNodePos, const int wordPos,
49             const NgramProperty *const ngramProperty, bool *const outAddedNewEntry);
50 
51     // TODO: Remove after stopping supporting v402.
52     // Remove an n-gram entry.
53     bool removeNgramEntry(const PtNodePosArrayView prevWordsPtNodePos, const int wordPos);
54 
55     // Add a shortcut target.
56     bool addShortcutTarget(const int wordPos, const CodePointArrayView targetCodePoints,
57             const int shortcutProbability);
58 
59  private:
60     DISALLOW_IMPLICIT_CONSTRUCTORS(DynamicPtUpdatingHelper);
61 
62     static const int CHILDREN_POSITION_FIELD_SIZE;
63 
64     BufferWithExtendableBuffer *const mBuffer;
65     const PtNodeReader *const mPtNodeReader;
66     PtNodeWriter *const mPtNodeWriter;
67 
68     bool createAndInsertNodeIntoPtNodeArray(const int parentPos,
69             const CodePointArrayView ptNodeCodePoints, const UnigramProperty *const unigramProperty,
70             int *const forwardLinkFieldPos);
71 
72     bool setPtNodeProbability(const PtNodeParams *const originalPtNodeParams,
73             const UnigramProperty *const unigramProperty, bool *const outAddedNewUnigram);
74 
75     bool createChildrenPtNodeArrayAndAChildPtNode(const PtNodeParams *const parentPtNodeParams,
76             const UnigramProperty *const unigramProperty,
77             const CodePointArrayView remainingCodePoints);
78 
79     bool createNewPtNodeArrayWithAChildPtNode(const int parentPos,
80             const CodePointArrayView ptNodeCodePoints,
81             const UnigramProperty *const unigramProperty);
82 
83     bool reallocatePtNodeAndAddNewPtNodes(const PtNodeParams *const reallocatingPtNodeParams,
84             const size_t overlappingCodePointCount, const UnigramProperty *const unigramProperty,
85             const CodePointArrayView newPtNodeCodePoints);
86 
87     const PtNodeParams getUpdatedPtNodeParams(const PtNodeParams *const originalPtNodeParams,
88             const bool isNotAWord, const bool isPossiblyOffensive, const bool isTerminal,
89             const int parentPos, const CodePointArrayView codePoints, const int probability) const;
90 
91     const PtNodeParams getPtNodeParamsForNewPtNode(const bool isNotAWord,
92             const bool isPossiblyOffensive, const bool isTerminal, const int parentPos,
93             const CodePointArrayView codePoints, const int probability) const;
94 };
95 } // namespace latinime
96 #endif /* LATINIME_DYNAMIC_PATRICIA_TRIE_UPDATING_HELPER_H */
97