Status LegacyReplyBuilder::addOutputDocs(DocumentRange docs) {
invariant(_state == State::kOutputDocs);
invariant(_allowAddingOutputDocs);
auto dataSize = docs.data().length();
auto hasSpace = _hasSpaceFor(dataSize);
if (!hasSpace.isOK()) {
return hasSpace;
}
// The temporary obj is used to address the case when where is not enough space.
// BSONArray overhead can not be estimated upfront.
std::vector<BSONObj> docsTmp{};
std::size_t lenTmp = 0;
std::size_t tmpIndex(_currentIndex);
for (auto&& it : docs) {
docsTmp.emplace_back(it.getOwned());
lenTmp += BSONObjBuilder::numStr(++tmpIndex).length() + 2; // space for storing array index
}
hasSpace = _hasSpaceFor(dataSize + lenTmp);
if (!hasSpace.isOK()) {
return hasSpace;
}
// vector::insert instead of swap allows to call addOutputDoc(s) multiple times
_outputDocs.insert(_outputDocs.end(),
std::make_move_iterator(docsTmp.begin()),
std::make_move_iterator(docsTmp.end()));
_currentIndex = tmpIndex;
_currentLength += lenTmp;
_currentLength += dataSize;
return Status::OK();
}
/**
* @brief analyze each word in a sentence
*
*/
vector<MorphologicalAnalyze> Dictionary::getMorphologicalAnalyze(const wstring & _sentence)
{
// tokenize sentence to words
vector<wstring> words = getDictionaryTools()->dictionarySplit(_sentence, WORD_DELIMITER);
// number of words
int n = (int) words.size();
// init morphological info list
vector<vector<MorphologicalInfo> > minfos = vector<vector<MorphologicalInfo> >();
vector<int> minfoSize = vector<int>();
// morphological analysis for each word
for (int i = 0; i < n; ++i)
{
// morphological analysis for words[i]
shared_ptr<vector<MorphologicalInfo> > minfo = std::make_shared<vector<MorphologicalInfo> >();
trie->getMorphologicalInfoList(words.at(i), minfo);
// if no MorphologicalInfo found
if (minfo->size() == 0)
{
// create MorphologicalInfo - NULL
MorphologicalInfo info;
info.featureListId = 0;
// then add to minfo
minfo->push_back(info);
}
minfos.push_back(*minfo);
minfoSize.push_back(minfo->size());
}
//init maxNGram;
int maxNGram = -1;
vector<int> maxNGramIndex;
// init zero-array length n
vector<int> tmpIndex(n, 0);
while (1)
{
// get combination of MorphologicalInfo
vector<MorphologicalInfo> _tmpMinfo = vector<MorphologicalInfo>();
for(int i = 0; i < n; ++i)
{
_tmpMinfo.push_back(minfos.at(i).at(tmpIndex.at(i)));
}
// get NGram value for the combination
int tmpNGramValue = this->getNGramTrie()->morphologicalNGramValue(_tmpMinfo);
// compare tmpNGramValue to maxNGram
if (tmpNGramValue > maxNGram)
{
maxNGram = tmpNGramValue;
maxNGramIndex = tmpIndex;
}
// change combination
int i = n - 1;
int z = 1;
while (i >= 0 && z == 1)
{
if (tmpIndex.at(i) + z < minfoSize.at(i))
{
tmpIndex.at(i) += z;
z = 0;
}
else
{
tmpIndex.at(i) = 0;
z = 1;
}
i--;
}
if (i < 0 && z == 1)
{
break;
}
}
// get result
//wcout << "maxNGram = " << maxNGram << endl;
vector<MorphologicalAnalyze> result = vector<MorphologicalAnalyze>();
for (int i = 0; i < n; ++i)
{
MorphologicalAnalyze mAnalyze;
mAnalyze.minfos = minfos.at(i);
mAnalyze.bestVariant = maxNGramIndex.at(i);
result.push_back(mAnalyze);
}
return result;
}
