void LTRTextBuilder::updateLTR_TokenFromVertex(
const LinguisticGraphVertex& vertex,
const LinguisticGraph& graph,
LTR_Token* tokenRep,
uint64_t offset) const {
// get data from the result of the linguistic analysis
Token* fullToken = get(vertex_token, graph, vertex);
MorphoSyntacticData* data = get(vertex_data, graph, vertex);
const FsaStringsPool& sp = (Common::MediaticData::MediaticData::single().stringsPool(m_language));
sort(data->begin(),data->end(),ltNormProperty(m_macroAccessor));
StringsPoolIndex norm(0),lastNorm(0);
LinguisticCode macro(0),lastMacro(0);
for (MorphoSyntacticData::const_iterator elemItr=data->begin();
elemItr!=data->end(); elemItr++) {
norm = elemItr->normalizedForm;
macro = m_macroAccessor->readValue(elemItr->properties);
if (norm == lastNorm && macro == lastMacro) {
continue;
}
else {
lastNorm=norm;
lastMacro=macro;
LimaString normStr= sp[norm];
// test if the same word was not already met at this position
bool selectionFlag = true;
LTR_Token::const_iterator itTok = tokenRep->begin();
while (selectionFlag && (itTok != tokenRep->end())) {
selectionFlag = (itTok->first->getLemma() != normStr) ||
(itTok->first->getCategory() != macro);
itTok ++;
}
if (selectionFlag) {
// test if the current token is a plain word
bool plainWordFlag =
this->isWordToSelect(normStr, macro,m_microAccessor->readValue(elemItr->properties));
BoWToken* bowToken = new BoWToken(normStr,macro,
fullToken->position() + offset,
fullToken->length());
bowToken->setInflectedForm(fullToken->stringForm());
tokenRep->push_back(make_pair(bowToken, plainWordFlag));
}
}
}
}
