SyntacticAnalysisXmlLogger.h

Go to the documentation of this file.

// Copyright 2002-2013 CEA LIST
// SPDX-FileCopyrightText: 2022 CEA LIST <gael.de-chalendar@cea.fr>
//
// SPDX-License-Identifier: MIT
 
#ifndef SYNTACTICANALYSISXMLLOGGERSPECIFICENTITIESXMLLOGGER_H
#define SYNTACTICANALYSISXMLLOGGERSPECIFICENTITIESXMLLOGGER_H
 
#include "SyntacticAnalysisExport.h"
#include "linguisticProcessing/core/LinguisticProcessors/AbstractTextualAnalysisDumper.h"
#include "linguisticProcessing/core/LinguisticAnalysisStructure/LinguisticGraph.h"
#include "linguisticProcessing/core/LinguisticAnalysisStructure/AnalysisGraph.h"
#include "linguisticProcessing/core/LinguisticAnalysisStructure/Token.h"
 
 
namespace Lima
{
namespace LinguisticProcessing
{
namespace SyntacticAnalysis
{
 
class SyntacticData;
 
#define SYNTACTICANALYSISXMLLOGGER_CLASSID "SyntacticAnalysisXmlLogger"
 
class LIMA_SYNTACTICANALYSIS_EXPORT SyntacticAnalysisXmlLogger : public AbstractTextualAnalysisDumper
{
public:
  SyntacticAnalysisXmlLogger();
 
  virtual ~SyntacticAnalysisXmlLogger();
 
  virtual void init(
    Common::XMLConfigurationFiles::GroupConfigurationStructure& unitConfiguration,
    Manager* manager) override
  ;
 
  virtual LimaStatusCode process(
    AnalysisContent& analysis) const override;
 
private:
  
  void dumpLimaData(std::ostream& os,
                  const LinguisticGraphVertex begin,
                  const LinguisticGraphVertex end,
                  const LinguisticAnalysisStructure::AnalysisGraph* pit,
                  const SyntacticAnalysis::SyntacticData* syntacticData,
                  const uint64_t offsetBegin=0) const;
  class DumpGraphVisitor : public boost::default_bfs_visitor
  {
    const SyntacticAnalysisXmlLogger& m_dumper;
    std::ostream& m_os;
    LinguisticGraphVertex m_lastVertex;
    const SyntacticAnalysis::SyntacticData* m_syntacticData;
    MediaId m_language;
    uint64_t m_offsetBegin;
    std::map< LinguisticAnalysisStructure::Token*, uint64_t >& m_tokens;
    std::vector< bool >& m_alreadyDumpedTokens;
    
    public:
    DumpGraphVisitor(const SyntacticAnalysisXmlLogger& dumper,
                    std::ostream& outputStream,
                    const LinguisticGraphVertex& end,
                    const SyntacticAnalysis::SyntacticData* syntacticData,
                    std::map< LinguisticAnalysisStructure::Token*, uint64_t >& tokens,
                    std::vector< bool >& alreadyDumpedTokens,
                    MediaId language,
                    const uint64_t offsetBegin):
          m_dumper(dumper),
          m_os(outputStream), 
          m_lastVertex(end), 
          m_syntacticData(syntacticData),
    m_language(language), 
    m_offsetBegin(offsetBegin), 
    m_tokens(tokens),
    m_alreadyDumpedTokens(alreadyDumpedTokens)
    {
    }
    
    void discover_vertex(LinguisticGraphVertex v, const LinguisticGraph& g);
    
    void examine_vertex(LinguisticGraphVertex v, const LinguisticGraph& g);
 
    void examine_edge(LinguisticGraphEdge e, const LinguisticGraph& g);
    
    class EndOfSearch : public std::runtime_error 
    {
    public:
        EndOfSearch():std::runtime_error("") {}
    };
  };
  
 
  static LimaString getPosition(const uint64_t position,
                                  const uint64_t offsetBegin);
  
  void outputVertex(const LinguisticGraphVertex v,
                    const LinguisticGraph& graph,
                    const uint64_t offsetBegin,
                    const SyntacticAnalysis::SyntacticData* syntacticData,
                    std::ostream& xmlStream,
                    std::map< LinguisticAnalysisStructure::Token*, uint64_t >& fullTokens,
                    std::vector< bool >& alreadyDumpedFullTokens) const;
                    
  void outputEdge(const LinguisticGraphEdge e,
                  const LinguisticGraph& graph,
                  std::ostream& xmlStream) const;
  
  MediaId m_language;
  
  const Common::PropertyCode::PropertyCodeManager* m_propertyCodeManager;
 
};
 
} // SyntacticAnalysis
} // LinguisticProcessing
} // Lima
 
#endif