LTRTextBuilder.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 LIMA_LINGUISTICPROCESSING_LTRTEXTBUILDER_H
#define LIMA_LINGUISTICPROCESSING_LTRTEXTBUILDER_H
 
#include "AnalysisDumpersExport.h"
#include "StopList.h"
 
//#include "linguisticProcessing/common/BagOfWords/bowText.h"
#include "linguisticProcessing/common/linearTextRepresentation/ltrToken.h"
#include "linguisticProcessing/common/linearTextRepresentation/ltrText.h"
#include "linguisticProcessing/core/LinguisticAnalysisStructure/AnalysisGraph.h"
#include "linguisticProcessing/core/LinguisticAnalysisStructure/LinguisticGraph.h"
#include "linguisticProcessing/core/TextSegmentation/SegmentationData.h"
 
#include <list>
#include <set>
#include <deque>
 
 
namespace Lima {
namespace LinguisticProcessing {
namespace AnalysisDumpers {
 
 
class LIMA_ANALYSISDUMPERS_EXPORT LTRTextBuilder {
 
// type definitions
// ----------------
private:
    typedef std::list<LinguisticGraphVertex> VERTICES_TO_EXPLORE_T;
    typedef std::set<LinguisticGraphVertex> EXPLORED_VERTICES_T;
 
 
// data
// ----
private:
    VERTICES_TO_EXPLORE_T m_verticesToExplore;
    EXPLORED_VERTICES_T m_exploredVertices;
    uint64_t m_currentOffset;
    MediaId m_language;
    std::shared_ptr<StopList> m_stopList;
    const Common::PropertyCode::PropertyAccessor* m_macroAccessor;
    const Common::PropertyCode::PropertyAccessor* m_microAccessor;
 
 
// methods
// -------
public:
    LTRTextBuilder(
    const MediaId& language,
        std::shared_ptr<StopList> stopList);
    void buildLTRTextFrom(
        const LinguisticGraph& graph,
        Lima::LinguisticProcessing::SegmentationData* sb,
        const LinguisticGraphVertex& graphFirstVertex,
        const LinguisticGraphVertex& graphLastVertex,
        boost::shared_ptr<Lima::Common::BagOfWords::LTR_Text> textRep,
        uint64_t offset);
 
 
    bool isWordToSelect(
        const Lima::LimaString& lemma,
        LinguisticCode macroCategory) const;
    bool isWordToSelect(
        const Lima::LimaString& lemma,
        LinguisticCode macroCategory,
        LinguisticCode microCategory) const;
    //}@
 
private:
    void addTokensToLTRTextFrom(
        const LinguisticGraph& graph,
        const LinguisticGraphVertex& firstVertex,
        const LinguisticGraphVertex& lastVertex,
        const LinguisticGraphVertex& graphLastVertex,
        boost::shared_ptr<Lima::Common::BagOfWords::LTR_Text> textRep,
        uint64_t offset,
        uint64_t* tokenCounter);
    void exploreVerticesFrom(
        const LinguisticGraphVertex& vertex,
        const LinguisticGraphVertex& graphLastVertex,
        const LinguisticGraph& graph);
    void updateLTR_TokenFromVertex(
        const LinguisticGraphVertex& vertex,
        const LinguisticGraph& graph,
        boost::shared_ptr<Lima::Common::BagOfWords::LTR_Token> tokenRep,
        uint64_t offset) const;
};
 
 
} // AnalysisDumpers
} // LinguisticProcessing
} // Lima
 
 
 
#endif  // LIMA_LINGUISTICPROCESSING_LTRTEXTBUILDER_H