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));
}
}
}
}
void DumpXMLVisitor::discover_vertex(LinguisticGraphVertex v,
const LinguisticGraph& g)
{
m_ostream << "<vertex id=\"" << v << "\">" << std::endl;
Token* t = get(vertex_token,g,v);
if (t != 0)
{
m_ostream << " <token>" << std::endl;
t->outputXml(m_ostream,m_propertyCodeManager,m_stringsPool);
m_ostream << " </token>" << std::endl;
}
MorphoSyntacticData* data = get(vertex_data,g,v);
if (data != 0 )
{
data->outputXml(m_ostream,m_propertyCodeManager,m_stringsPool);
}
m_ostream << "</vertex>" << std::endl;
}
void SyntacticAnalysisXmlLogger::outputVertex(const LinguisticGraphVertex v,
const LinguisticGraph& graph,
const uint64_t offsetBegin,
const SyntacticData* syntacticData,
std::ostream& xmlStream,
std::map< LinguisticAnalysisStructure::Token*, uint64_t >& tokens,
std::vector< bool >& alreadyDumpedTokens) const
{
if (v == syntacticData->iterator()->firstVertex() ||
v == syntacticData->iterator()->lastVertex())
{
xmlStream << "<vertex id=\"" << v << "\" />" << std::endl;
return;
}
Token* token = get(vertex_token, graph, v);
uint64_t tokenId = (*(tokens.find(token))).second;
// bool alreadyDumped = alreadyDumpedTokens[tokenId];
xmlStream << "<vertex id=\"" << v << "\" form=\"" << limastring2utf8stdstring(token->stringForm()) << "\" pos=\"" << getPosition(token->position(),offsetBegin) << "\" ";
const VertexChainIdProp& chains = get(vertex_chain_id, graph,v);
xmlStream << " >" << std::endl;
if (chains.size() > 0)
{
xmlStream << "<chains>" << std::endl;
VertexChainIdProp::const_iterator itChains, itChains_end;
itChains = chains.begin();
itChains_end = chains.end();
for (; itChains != itChains_end; itChains++)
{
const ChainIdStruct& ids = (*itChains);
xmlStream << "<chain type=\"";
if (ids.chainType() == Common::MediaticData::NO_CHAIN_TYPE)
xmlStream << "0";
else if (ids.chainType() == Common::MediaticData::NOMINAL)
xmlStream << "N";
else
xmlStream << "V";
xmlStream << "\" id=\"" << (ids.chainId()) << "\" />" << std::endl;
}
xmlStream << "</chains>" << std::endl;
}
const DependencyGraph* depGraph = syntacticData->dependencyGraph();
DependencyGraphVertex depV = syntacticData->depVertexForTokenVertex(v);
if (out_degree(depV, *depGraph) > 0)
{
xmlStream << "<dependents>" << std::endl;
DependencyGraphOutEdgeIt depIt, depIt_end;
boost::tie(depIt, depIt_end) = out_edges(depV, *depGraph);
for (; depIt != depIt_end; depIt++)
{
DependencyGraphVertex depTargV = target(*depIt, *depGraph);
LinguisticGraphVertex targV = syntacticData-> tokenVertexForDepVertex(depTargV);
// CEdgeDepChainIdPropertyMap chainsMap = get(edge_depchain_id, *depGraph);
CEdgeDepRelTypePropertyMap relTypeMap = get(edge_deprel_type, *depGraph);
xmlStream << "<dep v=\"" << targV;
// xmlStream << "\" c=\"" << chainsMap[*depIt];
std::string relName=static_cast<const Common::MediaticData::LanguageData&>(Common::MediaticData::MediaticData::single().mediaData(m_language)).getSyntacticRelationName(relTypeMap[*depIt]);
if (relName.empty())
{
relName="UNKNOWN";
}
xmlStream << "\" t=\"" << relName << "\" />" << std::endl;
}
xmlStream << "</dependents>" << std::endl;
}
const FsaStringsPool& sp=Common::MediaticData::MediaticData::single().stringsPool(m_language);
MorphoSyntacticData* word = get(vertex_data, graph, v);
word->outputXml(xmlStream,*m_propertyCodeManager,sp);
xmlStream << "<ref>" << tokenId << "</ref>" << std::endl;
alreadyDumpedTokens[tokenId] = true;
xmlStream << "</vertex>" << std::endl;
}
