Home | History | Annotate | Download | only in pt_common 
      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_PATRICIA_TRIE_READING_UTILS_H
     18 #define LATINIME_PATRICIA_TRIE_READING_UTILS_H
     19 
     20 #include <cstdint>
     21 
     22 #include "defines.h"
     23 
     24 namespace latinime {
     25 
     26 class DictionaryShortcutsStructurePolicy;
     27 class DictionaryBigramsStructurePolicy;
     28 
     29 class PatriciaTrieReadingUtils {
     30  public:
     31     typedef uint8_t NodeFlags;
     32 
     33     static int getPtNodeArraySizeAndAdvancePosition(const uint8_t *const buffer, int *const pos);
     34 
     35     static NodeFlags getFlagsAndAdvancePosition(const uint8_t *const buffer, int *const pos);
     36 
     37     static int getCodePointAndAdvancePosition(const uint8_t *const buffer, int *const pos);
     38 
     39     // Returns the number of read characters.
     40     static int getCharsAndAdvancePosition(const uint8_t *const buffer, const NodeFlags flags,
     41             const int maxLength, int *const outBuffer, int *const pos);
     42 
     43     // Returns the number of skipped characters.
     44     static int skipCharacters(const uint8_t *const buffer, const NodeFlags flags,
     45             const int maxLength, int *const pos);
     46 
     47     static int readProbabilityAndAdvancePosition(const uint8_t *const buffer, int *const pos);
     48 
     49     static int readChildrenPositionAndAdvancePosition(const uint8_t *const buffer,
     50             const NodeFlags flags, int *const pos);
     51 
     52     /**
     53      * Node Flags
     54      */
     55     static AK_FORCE_INLINE bool isBlacklisted(const NodeFlags flags) {
     56         return (flags & FLAG_IS_BLACKLISTED) != 0;
     57     }
     58 
     59     static AK_FORCE_INLINE bool isNotAWord(const NodeFlags flags) {
     60         return (flags & FLAG_IS_NOT_A_WORD) != 0;
     61     }
     62 
     63     static AK_FORCE_INLINE bool isTerminal(const NodeFlags flags) {
     64         return (flags & FLAG_IS_TERMINAL) != 0;
     65     }
     66 
     67     static AK_FORCE_INLINE bool hasShortcutTargets(const NodeFlags flags) {
     68         return (flags & FLAG_HAS_SHORTCUT_TARGETS) != 0;
     69     }
     70 
     71     static AK_FORCE_INLINE bool hasBigrams(const NodeFlags flags) {
     72         return (flags & FLAG_HAS_BIGRAMS) != 0;
     73     }
     74 
     75     static AK_FORCE_INLINE bool hasMultipleChars(const NodeFlags flags) {
     76         return (flags & FLAG_HAS_MULTIPLE_CHARS) != 0;
     77     }
     78 
     79     static AK_FORCE_INLINE bool hasChildrenInFlags(const NodeFlags flags) {
     80         return FLAG_CHILDREN_POSITION_TYPE_NOPOSITION != (MASK_CHILDREN_POSITION_TYPE & flags);
     81     }
     82 
     83     static AK_FORCE_INLINE NodeFlags createAndGetFlags(const bool isBlacklisted,
     84             const bool isNotAWord, const bool isTerminal, const bool hasShortcutTargets,
     85             const bool hasBigrams, const bool hasMultipleChars,
     86             const int childrenPositionFieldSize) {
     87         NodeFlags nodeFlags = 0;
     88         nodeFlags = isBlacklisted ? (nodeFlags | FLAG_IS_BLACKLISTED) : nodeFlags;
     89         nodeFlags = isNotAWord ? (nodeFlags | FLAG_IS_NOT_A_WORD) : nodeFlags;
     90         nodeFlags = isTerminal ? (nodeFlags | FLAG_IS_TERMINAL) : nodeFlags;
     91         nodeFlags = hasShortcutTargets ? (nodeFlags | FLAG_HAS_SHORTCUT_TARGETS) : nodeFlags;
     92         nodeFlags = hasBigrams ? (nodeFlags | FLAG_HAS_BIGRAMS) : nodeFlags;
     93         nodeFlags = hasMultipleChars ? (nodeFlags | FLAG_HAS_MULTIPLE_CHARS) : nodeFlags;
     94         if (childrenPositionFieldSize == 1) {
     95             nodeFlags |= FLAG_CHILDREN_POSITION_TYPE_ONEBYTE;
     96         } else if (childrenPositionFieldSize == 2) {
     97             nodeFlags |= FLAG_CHILDREN_POSITION_TYPE_TWOBYTES;
     98         } else if (childrenPositionFieldSize == 3) {
     99             nodeFlags |= FLAG_CHILDREN_POSITION_TYPE_THREEBYTES;
    100         } else {
    101             nodeFlags |= FLAG_CHILDREN_POSITION_TYPE_NOPOSITION;
    102         }
    103         return nodeFlags;
    104     }
    105 
    106     static void readPtNodeInfo(const uint8_t *const dictBuf, const int ptNodePos,
    107             const DictionaryShortcutsStructurePolicy *const shortcutPolicy,
    108             const DictionaryBigramsStructurePolicy *const bigramPolicy,
    109             NodeFlags *const outFlags, int *const outCodePointCount, int *const outCodePoint,
    110             int *const outProbability, int *const outChildrenPos, int *const outShortcutPos,
    111             int *const outBigramPos, int *const outSiblingPos);
    112 
    113  private:
    114     DISALLOW_IMPLICIT_CONSTRUCTORS(PatriciaTrieReadingUtils);
    115 
    116     static const NodeFlags MASK_CHILDREN_POSITION_TYPE;
    117     static const NodeFlags FLAG_CHILDREN_POSITION_TYPE_NOPOSITION;
    118     static const NodeFlags FLAG_CHILDREN_POSITION_TYPE_ONEBYTE;
    119     static const NodeFlags FLAG_CHILDREN_POSITION_TYPE_TWOBYTES;
    120     static const NodeFlags FLAG_CHILDREN_POSITION_TYPE_THREEBYTES;
    121 
    122     static const NodeFlags FLAG_HAS_MULTIPLE_CHARS;
    123     static const NodeFlags FLAG_IS_TERMINAL;
    124     static const NodeFlags FLAG_HAS_SHORTCUT_TARGETS;
    125     static const NodeFlags FLAG_HAS_BIGRAMS;
    126     static const NodeFlags FLAG_IS_NOT_A_WORD;
    127     static const NodeFlags FLAG_IS_BLACKLISTED;
    128 };
    129 } // namespace latinime
    130 #endif /* LATINIME_PATRICIA_TRIE_NODE_READING_UTILS_H */
    131