C++ AnnotationData::annotation方法代码示例

bool EntityGroupTransition::
compare(const LinguisticAnalysisStructure::AnalysisGraph& graph,
        const LinguisticGraphVertex& v,
        AnalysisContent& analysis,
        const LinguisticAnalysisStructure::Token* /*token*/,
        const LinguisticAnalysisStructure::MorphoSyntacticData* /*data*/) const
{
  // should compare to vertex ?
  AnnotationData* annotationData = static_cast< AnnotationData* >(analysis.getData("AnnotationData"));
  if (annotationData==0) {
    AULOGINIT;
    LDEBUG << "EntityGroupTransition::compare: no annotation graph available !";
    return false;
  }

  // find annotationGraphVertex matching the vertex of the current graph
  std::set<AnnotationGraphVertex> matches = annotationData->matches(graph.getGraphId(), v, "annot");
  if (matches.empty())
  {
    AULOGINIT;
    LDEBUG << "annotation ("<<graph.getGraphId()<<", "<<v<<", \"annot\") available";
    return false;
  }
  AnnotationGraphVertex annotVertex = *(matches.begin());

  if (!annotationData->hasAnnotation(annotVertex, m_entityAnnotation))
  {
    AULOGINIT;
    LDEBUG << "EntityGroupTransition::compare: No " << m_entityAnnotation << " annotation available on " << v;
    return false;
  }
  
  const SpecificEntityAnnotation* se =
    annotationData->annotation(annotVertex, m_entityAnnotation).
    pointerValue<SpecificEntityAnnotation>();
  Common::MediaticData::EntityType type = se->getType();
  AULOGINIT;
  LDEBUG << "EntityGroupTransition::compare: type = " << type << ", groupId = " << type.getGroupId();
  LDEBUG << "EntityGroupTransition::compare: m_entityGroupId = " << m_entityGroupId;
  LDEBUG << "EntityGroupTransition::compare: tests m_entityGroupId == type.getGroupId() = " << (m_entityGroupId == type.getGroupId());
  return( m_entityGroupId == type.getGroupId() );
}

LimaStatusCode SpecificEntitiesXmlLogger::process(
  AnalysisContent& analysis) const
{
  SELOGINIT;
  LDEBUG << "SpecificEntitiesXmlLogger::process";
  TimeUtils::updateCurrentTime();

  AnnotationData* annotationData = static_cast< AnnotationData* >(analysis.getData("AnnotationData"));
  if (annotationData == 0) {
    SELOGINIT;
    LERROR << "no annotationData ! abort";
    return MISSING_DATA;
  }
  
  
  LinguisticAnalysisStructure::AnalysisGraph* graphp = static_cast<LinguisticAnalysisStructure::AnalysisGraph*>(analysis.getData(m_graph));
  if (graphp == 0) {
    SELOGINIT;
    LERROR << "no graph "<< m_graph <<" ! abort";
    return MISSING_DATA;
  }
  const LinguisticAnalysisStructure::AnalysisGraph& graph = *graphp;
  LinguisticGraph* lingGraph = const_cast<LinguisticGraph*>(graph.getGraph());
  VertexTokenPropertyMap tokenMap = get(vertex_token, *lingGraph);
  
  LinguisticMetaData* metadata=static_cast<LinguisticMetaData*>(analysis.getData("LinguisticMetaData"));
  if (metadata == 0) {
      SELOGINIT;
      LERROR << "no LinguisticMetaData ! abort";
      return MISSING_DATA;
  }

  DumperStream* dstream=initialize(analysis);
  ostream& out=dstream->out();

  uint64_t offset(0);
  try {
    offset=atoi(metadata->getMetaData("StartOffset").c_str());
  }
  catch (LinguisticProcessingException& ) {
    // do nothing: not set in analyzeText (only in analyzeXmlDocuments)
  }

  uint64_t offsetIndexingNode(0);
  try {
    offsetIndexingNode=atoi(metadata->getMetaData("StartOffsetIndexingNode").c_str());
  }
  catch (LinguisticProcessingException& ) {
    // do nothing: not set in analyzeText (only in analyzeXmlDocuments)
  }

  std::string docId("");
  try {
    docId=metadata->getMetaData("DocId");
  }
  catch (LinguisticProcessingException& ) {
    // do nothing: not set in analyzeText (only in analyzeXmlDocuments)
  }

  if (m_compactFormat) {
    out << "<entities docid=\"" << docId
    << "\" offsetNode=\"" << offsetIndexingNode 
    << "\" offset=\"" << offset
    << "\">" << endl;
  }
  else {
    out << "<specific_entities>" << endl;
  }
//   SELOGINIT;

  if (m_followGraph) {
    // instead of looking to all annotations, follow the graph (in
    // morphological graph, some vertices are not related to main graph:
    // idiomatic expressions parts and named entity parts)
    // -> this will not include nested entities

    AnalysisGraph* tokenList=static_cast<AnalysisGraph*>(analysis.getData(m_graph));
    if (tokenList==0) {
      LERROR << "graph " << m_graph << " has not been produced: check pipeline";
      return MISSING_DATA;
    }
    LinguisticGraph* graph=tokenList->getGraph();
    //const FsaStringsPool& sp=Common::MediaticData::MediaticData::single().stringsPool(m_language);
    
    std::queue<LinguisticGraphVertex> toVisit;
    std::set<LinguisticGraphVertex> visited;
    toVisit.push(tokenList->firstVertex());
    
    LinguisticGraphOutEdgeIt outItr,outItrEnd;
    while (!toVisit.empty()) {
      LinguisticGraphVertex v=toVisit.front();
      toVisit.pop();
      if (v == tokenList->lastVertex()) {
        continue;
      }
      
      for (boost::tie(outItr,outItrEnd)=out_edges(v,*graph); outItr!=outItrEnd; outItr++) 
      {
        LinguisticGraphVertex next=target(*outItr,*graph);
        if (visited.find(next)==visited.end())
//.........这里部分代码省略.........

      LimaStatusCode CorefSolvingNormalizedXmlLogger::process(
        AnalysisContent& analysis) const
      {
//         COREFSOLVERLOGINIT;
        TimeUtils::updateCurrentTime();
        AnnotationData* annotationData = static_cast<AnnotationData*>(analysis.getData("AnnotationData"));
        const LinguisticAnalysisStructure::AnalysisGraph& graph = *(static_cast<LinguisticAnalysisStructure::AnalysisGraph*>(analysis.getData(m_graph)));

//         LinguisticGraph* lingGraph = const_cast<LinguisticGraph*>(graph.getGraph());
        LinguisticMetaData* metadata=static_cast<LinguisticMetaData*>(analysis.getData("LinguisticMetaData"));
        if (metadata == 0)
        {
          COREFSOLVERLOGINIT;
          LERROR << "no LinguisticMetaData ! abort" << LENDL;
          return MISSING_DATA;
        }

        ofstream out;
        if (!openLogFile(out,metadata->getMetaData("FileName")))
        {
          COREFSOLVERLOGINIT;
          LERROR << "Can't open log file " << LENDL;
          return UNKNOWN_ERROR;
        }

        out << "<coreferences>" << endl;


        //   LDEBUG << "CorefSolvingNormalizedXmlLogger on graph " << m_graph << LENDL;
        AnnotationGraphVertexIt itv, itv_end;
        boost::tie(itv, itv_end) = vertices(annotationData->getGraph());
        for (; itv != itv_end; itv++)
        {
          // process
          //LDEBUG << "CorefSolvingNormalizedXmlLogger on annotation vertex " << *itv << LENDL;
          if (annotationData->hasAnnotation(*itv,utf8stdstring2limastring("Coreferent")))
            //if (annotationData->hasAnnotation(*itv,utf8stdstring2limastring("Coreferent")))
          {
            CoreferentAnnotation* annot ;
            try
            {
              annot = annotationData->annotation(*itv,utf8stdstring2limastring("Coreferent"))
                      .pointerValue<CoreferentAnnotation>();
            }
            catch (const boost::bad_any_cast& )
            {
              COREFSOLVERLOGINIT;
              LERROR << "One annotation on vertex " << *itv << " you are trying to cast is not a Coreference; Coreference not logged" << LENDL;
              for (int i = 0; i < 19 ; i++)
              {
                LERROR << "annot "<< i << " : " << limastring2utf8stdstring(annotationData->annotationName(i)) << LENDL ;
              }
              continue;
            }
            LinguisticProcessing::LinguisticAnalysisStructure::Token* token = get(vertex_token, *graph.getGraph(), annot->morphVertex());
            if (token == 0)
            {
              COREFSOLVERLOGINIT;
              LERROR << "Vertex " << *itv << " has no entry in the analysis graph token map. This should not happen !!" << LENDL;
            }
            else
            {
              CoreferentAnnotation* antecedent;
//               bool hasAntecedent = false;
              AnnotationGraphOutEdgeIt it, it_end;
              boost::tie(it, it_end) = boost::out_edges(static_cast<AnnotationGraphVertex>(*itv), annotationData->getGraph());

              for (; it != it_end; it++)
              {
                if (annotationData->hasAnnotation(target(*it,annotationData->getGraph()),utf8stdstring2limastring("Coreferent")))
                {
                  try
                  {
                    antecedent = annotationData->annotation(target(*it, annotationData->getGraph()), utf8stdstring2limastring("Coreferent")).pointerValue<CoreferentAnnotation>();
//                     hasAntecedent = true;
                  }
                  catch (const boost::bad_any_cast& )
                  {
                    COREFSOLVERLOGINIT;
                    LERROR << "One annotation on vertex you are trying to cast resulting from an edge out of " << *itv << " is not a Coreference; Coreference not logged" << LENDL;
                    continue;
                  }
                }
              }
              out << "  <reference>\n"
              << "    <pos>" << get(vertex_token,*graph.getGraph(),annot->morphVertex())->position() << "</pos>\n"
              << "    <len>" << token->stringForm().length() << "</len>\n"
              << "    <string>"<< limastring2utf8stdstring(transcodeToXmlEntities(token->stringForm())) << "</string>\n"
              << "    <npId>" << annot->id() << "</npId>\n"
              << "    <posVertex>" << annot->morphVertex() << "</posVertex>\n";
              //if (hasAntecedent)
              if (false)
              {
                out << "    <npRef>" << antecedent->id() << "</npRef>\n";
                out << "    <refPosVertex>" << antecedent->morphVertex() << "</refPosVertex>\n";
              }
              out << "    <categ>" << annot->categ() << "</categ>\n"
                    << "  </reference>\n"
              << endl;
            }
//.........这里部分代码省略.........