HyphenWordAlternatives.h

Go to the documentation of this file.

// Copyright 2002-2020 CEA LIST
// SPDX-FileCopyrightText: 2022 CEA LIST <gael.de-chalendar@cea.fr>
//
// SPDX-License-Identifier: MIT
 
#ifndef LIMA_MORPHOLOGICANALYSIS_HYPHENWORDALTERNATIVES_H
#define LIMA_MORPHOLOGICANALYSIS_HYPHENWORDALTERNATIVES_H
 
#include "MorphologicAnalysisExport.h"
#include "common/MediaProcessors/MediaProcessUnit.h"
#include "linguisticProcessing/core/FlatTokenizer/Tokenizer.h"
#include "linguisticProcessing/core/AnalysisDict/AbstractAnalysisDictionary.h"
#include "linguisticProcessing/core/FlatTokenizer/CharChart.h"
#include "linguisticProcessing/core/LinguisticAnalysisStructure/AnalysisGraph.h"
#include "linguisticProcessing/core/TextSegmentation/SegmentationData.h"
#include "AlternativesReader.h"
 
namespace Lima {
  namespace Common {
    namespace AnnotationGraphs {
      class AnnotationData;
    }
  }
namespace LinguisticProcessing {
namespace MorphologicAnalysis {
 
#define HYPHENWORDALTERNATIVESFACTORY_CLASSID "HyphenWordAlternatives"
 
class LIMA_MORPHOLOGICANALYSIS_EXPORT HyphenWordAlternatives : public MediaProcessUnit
{
 
public:
    HyphenWordAlternatives();
    virtual ~HyphenWordAlternatives();
 
  void init(
    Common::XMLConfigurationFiles::GroupConfigurationStructure& unitConfiguration,
    Manager* manager) override
  ;
 
  LimaStatusCode process(
    AnalysisContent& analysis) const override;
 
protected:
 
private:
 
    std::shared_ptr<FlatTokenizer::Tokenizer> m_tokenizer;
    std::shared_ptr<AnalysisDict::AbstractAnalysisDictionary> m_dictionary;
    std::shared_ptr<FlatTokenizer::CharChart> m_charChart;
    bool m_deleteHyphenWord;
    bool m_confidentMode;
    MediaId m_language;
    MediaId m_engLanguageId;
    std::shared_ptr<AlternativesReader> m_reader;
    std::string m_sentBoundariesName;
 
    void makeHyphenSplitAlternativeFor(
        LinguisticGraphVertex splitted,
        LinguisticGraph* graph,
        Common::AnnotationGraphs::AnnotationData* annotationData,
        SegmentationData* sb) const;
 
    bool isWorthSplitting(
        LinguisticGraphVertex splitted,
        LinguisticGraph* graph) const;
 
};
 
} // closing namespace MorphologicAnalysis
} // closing namespace LinguisticProcessing
} // closing namespace Lima
 
#endif // LIMA_MORPHOLOGICANALYSIS_HYPHENWORDALTERNATIVES_H