dic_traverse_session.cpp revision 4c2767857a02c9cf18a9579aa0391fd09b3fe411
13107b467c91c471ce4e00c5d8de559f7b0da2cd6Satoshi Kataoka/*
23107b467c91c471ce4e00c5d8de559f7b0da2cd6Satoshi Kataoka * Copyright (C) 2012 The Android Open Source Project
33107b467c91c471ce4e00c5d8de559f7b0da2cd6Satoshi Kataoka *
43107b467c91c471ce4e00c5d8de559f7b0da2cd6Satoshi Kataoka * Licensed under the Apache License, Version 2.0 (the "License");
53107b467c91c471ce4e00c5d8de559f7b0da2cd6Satoshi Kataoka * you may not use this file except in compliance with the License.
63107b467c91c471ce4e00c5d8de559f7b0da2cd6Satoshi Kataoka * You may obtain a copy of the License at
73107b467c91c471ce4e00c5d8de559f7b0da2cd6Satoshi Kataoka *
83107b467c91c471ce4e00c5d8de559f7b0da2cd6Satoshi Kataoka *      http://www.apache.org/licenses/LICENSE-2.0
93107b467c91c471ce4e00c5d8de559f7b0da2cd6Satoshi Kataoka *
103107b467c91c471ce4e00c5d8de559f7b0da2cd6Satoshi Kataoka * Unless required by applicable law or agreed to in writing, software
113107b467c91c471ce4e00c5d8de559f7b0da2cd6Satoshi Kataoka * distributed under the License is distributed on an "AS IS" BASIS,
123107b467c91c471ce4e00c5d8de559f7b0da2cd6Satoshi Kataoka * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
133107b467c91c471ce4e00c5d8de559f7b0da2cd6Satoshi Kataoka * See the License for the specific language governing permissions and
143107b467c91c471ce4e00c5d8de559f7b0da2cd6Satoshi Kataoka * limitations under the License.
153107b467c91c471ce4e00c5d8de559f7b0da2cd6Satoshi Kataoka */
163107b467c91c471ce4e00c5d8de559f7b0da2cd6Satoshi Kataoka
17b68e73448104714e8f12f89a1e00fb10b5fd14c4Ken Wakasa#include "suggest/core/session/dic_traverse_session.h"
18b68e73448104714e8f12f89a1e00fb10b5fd14c4Ken Wakasa
193107b467c91c471ce4e00c5d8de559f7b0da2cd6Satoshi Kataoka#include "defines.h"
20a65c267b1f1207e54c6f821148c600e3899b7f9cKen Wakasa#include "suggest/core/dictionary/dictionary.h"
2176e579c7caf2ef04f440be21c27377fe0b4150ffKeisuke Kuroyanagi#include "suggest/core/policy/dictionary_header_structure_policy.h"
22d81654cd61bd10f7cb56bfa4c89b34e9cfb18598Keisuke Kuroyanagi#include "suggest/core/policy/dictionary_structure_with_buffer_policy.h"
233107b467c91c471ce4e00c5d8de559f7b0da2cd6Satoshi Kataoka
243107b467c91c471ce4e00c5d8de559f7b0da2cd6Satoshi Kataokanamespace latinime {
253107b467c91c471ce4e00c5d8de559f7b0da2cd6Satoshi Kataoka
264c2767857a02c9cf18a9579aa0391fd09b3fe411Keisuke Kuroyanagi// 256K bytes threshold is heuristically used to distinguish dictionaries containing many unigrams
274c2767857a02c9cf18a9579aa0391fd09b3fe411Keisuke Kuroyanagi// (e.g. main dictionary) from small dictionaries (e.g. contacts...)
284c2767857a02c9cf18a9579aa0391fd09b3fe411Keisuke Kuroyanagiconst int DicTraverseSession::DICTIONARY_SIZE_THRESHOLD_TO_USE_LARGE_CACHE_FOR_SUGGESTION =
294c2767857a02c9cf18a9579aa0391fd09b3fe411Keisuke Kuroyanagi        256 * 1024;
304c2767857a02c9cf18a9579aa0391fd09b3fe411Keisuke Kuroyanagi
313107b467c91c471ce4e00c5d8de559f7b0da2cd6Satoshi Kataokavoid DicTraverseSession::init(const Dictionary *const dictionary, const int *prevWord,
32338ef65077f43d6a35665a5c8eb6a44928332547Keisuke Kuroyanagi        int prevWordLength, const SuggestOptions *const suggestOptions) {
333107b467c91c471ce4e00c5d8de559f7b0da2cd6Satoshi Kataoka    mDictionary = dictionary;
3476e579c7caf2ef04f440be21c27377fe0b4150ffKeisuke Kuroyanagi    mMultiWordCostMultiplier = getDictionaryStructurePolicy()->getHeaderStructurePolicy()
35d81654cd61bd10f7cb56bfa4c89b34e9cfb18598Keisuke Kuroyanagi            ->getMultiWordCostMultiplier();
36338ef65077f43d6a35665a5c8eb6a44928332547Keisuke Kuroyanagi    mSuggestOptions = suggestOptions;
373107b467c91c471ce4e00c5d8de559f7b0da2cd6Satoshi Kataoka    if (!prevWord) {
388a7129530b4dbb2453e88747f251d30a2ce75a45Keisuke Kuroynagi        mPrevWordPos = NOT_A_VALID_WORD_POS;
393107b467c91c471ce4e00c5d8de559f7b0da2cd6Satoshi Kataoka        return;
403107b467c91c471ce4e00c5d8de559f7b0da2cd6Satoshi Kataoka    }
415fa33a701d4b8405254d3369c714702481a90e6bTom Ouyang    // TODO: merge following similar calls to getTerminalPosition into one case-insensitive call.
42d81654cd61bd10f7cb56bfa4c89b34e9cfb18598Keisuke Kuroyanagi    mPrevWordPos = getDictionaryStructurePolicy()->getTerminalNodePositionOfWord(
43e1ebef6124241ef51d5ed17884e6299a330d496bKeisuke Kuroyanagi            prevWord, prevWordLength, false /* forceLowerCaseSearch */);
448a7129530b4dbb2453e88747f251d30a2ce75a45Keisuke Kuroynagi    if (mPrevWordPos == NOT_A_VALID_WORD_POS) {
455fa33a701d4b8405254d3369c714702481a90e6bTom Ouyang        // Check bigrams for lower-cased previous word if original was not found. Useful for
465fa33a701d4b8405254d3369c714702481a90e6bTom Ouyang        // auto-capitalized words like "The [current_word]".
47d81654cd61bd10f7cb56bfa4c89b34e9cfb18598Keisuke Kuroyanagi        mPrevWordPos = getDictionaryStructurePolicy()->getTerminalNodePositionOfWord(
48e1ebef6124241ef51d5ed17884e6299a330d496bKeisuke Kuroyanagi                prevWord, prevWordLength, true /* forceLowerCaseSearch */);
495fa33a701d4b8405254d3369c714702481a90e6bTom Ouyang    }
503107b467c91c471ce4e00c5d8de559f7b0da2cd6Satoshi Kataoka}
513107b467c91c471ce4e00c5d8de559f7b0da2cd6Satoshi Kataoka
523107b467c91c471ce4e00c5d8de559f7b0da2cd6Satoshi Kataokavoid DicTraverseSession::setupForGetSuggestions(const ProximityInfo *pInfo,
533107b467c91c471ce4e00c5d8de559f7b0da2cd6Satoshi Kataoka        const int *inputCodePoints, const int inputSize, const int *const inputXs,
543107b467c91c471ce4e00c5d8de559f7b0da2cd6Satoshi Kataoka        const int *const inputYs, const int *const times, const int *const pointerIds,
553107b467c91c471ce4e00c5d8de559f7b0da2cd6Satoshi Kataoka        const float maxSpatialDistance, const int maxPointerCount) {
563107b467c91c471ce4e00c5d8de559f7b0da2cd6Satoshi Kataoka    mProximityInfo = pInfo;
573107b467c91c471ce4e00c5d8de559f7b0da2cd6Satoshi Kataoka    mMaxPointerCount = maxPointerCount;
583107b467c91c471ce4e00c5d8de559f7b0da2cd6Satoshi Kataoka    initializeProximityInfoStates(inputCodePoints, inputXs, inputYs, times, pointerIds, inputSize,
593107b467c91c471ce4e00c5d8de559f7b0da2cd6Satoshi Kataoka            maxSpatialDistance, maxPointerCount);
603107b467c91c471ce4e00c5d8de559f7b0da2cd6Satoshi Kataoka}
613107b467c91c471ce4e00c5d8de559f7b0da2cd6Satoshi Kataoka
62d81654cd61bd10f7cb56bfa4c89b34e9cfb18598Keisuke Kuroyanagiconst DictionaryStructureWithBufferPolicy *DicTraverseSession::getDictionaryStructurePolicy()
63d81654cd61bd10f7cb56bfa4c89b34e9cfb18598Keisuke Kuroyanagi        const {
64d81654cd61bd10f7cb56bfa4c89b34e9cfb18598Keisuke Kuroyanagi    return mDictionary->getDictionaryStructurePolicy();
653107b467c91c471ce4e00c5d8de559f7b0da2cd6Satoshi Kataoka}
663107b467c91c471ce4e00c5d8de559f7b0da2cd6Satoshi Kataoka
6780ca7abea32a97acefcd8a8cb6145f0cdc8f0503Keisuke Kuroyanagivoid DicTraverseSession::resetCache(const int thresholdForNextActiveDicNodes, const int maxWords) {
6880ca7abea32a97acefcd8a8cb6145f0cdc8f0503Keisuke Kuroyanagi    mDicNodesCache.reset(thresholdForNextActiveDicNodes /* nextActiveSize */,
6980ca7abea32a97acefcd8a8cb6145f0cdc8f0503Keisuke Kuroyanagi            maxWords /* terminalSize */);
709559dd2e30de288a9ff7069bfc59f8500b949a88Tom Ouyang    mMultiBigramMap.clear();
713107b467c91c471ce4e00c5d8de559f7b0da2cd6Satoshi Kataoka    mPartiallyCommited = false;
723107b467c91c471ce4e00c5d8de559f7b0da2cd6Satoshi Kataoka}
733107b467c91c471ce4e00c5d8de559f7b0da2cd6Satoshi Kataoka
743107b467c91c471ce4e00c5d8de559f7b0da2cd6Satoshi Kataokavoid DicTraverseSession::initializeProximityInfoStates(const int *const inputCodePoints,
753107b467c91c471ce4e00c5d8de559f7b0da2cd6Satoshi Kataoka        const int *const inputXs, const int *const inputYs, const int *const times,
763107b467c91c471ce4e00c5d8de559f7b0da2cd6Satoshi Kataoka        const int *const pointerIds, const int inputSize, const float maxSpatialDistance,
773107b467c91c471ce4e00c5d8de559f7b0da2cd6Satoshi Kataoka        const int maxPointerCount) {
783107b467c91c471ce4e00c5d8de559f7b0da2cd6Satoshi Kataoka    ASSERT(1 <= maxPointerCount && maxPointerCount <= MAX_POINTER_COUNT_G);
793107b467c91c471ce4e00c5d8de559f7b0da2cd6Satoshi Kataoka    mInputSize = 0;
803107b467c91c471ce4e00c5d8de559f7b0da2cd6Satoshi Kataoka    for (int i = 0; i < maxPointerCount; ++i) {
813107b467c91c471ce4e00c5d8de559f7b0da2cd6Satoshi Kataoka        mProximityInfoStates[i].initInputParams(i, maxSpatialDistance, getProximityInfo(),
823107b467c91c471ce4e00c5d8de559f7b0da2cd6Satoshi Kataoka                inputCodePoints, inputSize, inputXs, inputYs, times, pointerIds,
833107b467c91c471ce4e00c5d8de559f7b0da2cd6Satoshi Kataoka                maxPointerCount == MAX_POINTER_COUNT_G
843107b467c91c471ce4e00c5d8de559f7b0da2cd6Satoshi Kataoka                /* TODO: this is a hack. fix proximity info state */);
853107b467c91c471ce4e00c5d8de559f7b0da2cd6Satoshi Kataoka        mInputSize += mProximityInfoStates[i].size();
863107b467c91c471ce4e00c5d8de559f7b0da2cd6Satoshi Kataoka    }
873107b467c91c471ce4e00c5d8de559f7b0da2cd6Satoshi Kataoka}
883107b467c91c471ce4e00c5d8de559f7b0da2cd6Satoshi Kataoka} // namespace latinime
89