IDRipper.h

Go to the documentation of this file.

// Copyright (c) 2002-2023, The OpenMS Team -- EKU Tuebingen, ETH Zurich, and FU Berlin
// SPDX-License-Identifier: BSD-3-Clause
//
// --------------------------------------------------------------------------
// $Maintainer: Timo Sachsenberg$
// $Authors: Immanuel Luhn, Leon Kuchenbecker$
// --------------------------------------------------------------------------
#pragma once
 
#include <OpenMS/DATASTRUCTURES/DefaultParamHandler.h>
#include <OpenMS/METADATA/PeptideIdentification.h>
#include <OpenMS/METADATA/ProteinIdentification.h>
#include <unordered_map>
 
 
namespace OpenMS
{
  class OPENMS_DLLAPI IDRipper :
    public DefaultParamHandler
  {
public:
    enum OriginAnnotationFormat { FILE_ORIGIN = 0, MAP_INDEX = 1, ID_MERGE_INDEX = 2, UNKNOWN_OAF = 3, SIZE_OF_ORIGIN_ANNOTATION_FORMAT = 4 };
 
    static const std::array<std::string, SIZE_OF_ORIGIN_ANNOTATION_FORMAT> names_of_OriginAnnotationFormat;
 
    struct OPENMS_DLLAPI IdentificationRuns
    {
        std::map<String, UInt> index_map;
        std::vector<StringList> spectra_data;
 
        IdentificationRuns(const std::vector<ProteinIdentification>& prot_ids);
    };
 
    struct OPENMS_DLLAPI RipFileIdentifier
    {
        UInt ident_run_idx{};
        UInt file_origin_idx{};
        String out_basename;
        String origin_fullname;
 
        RipFileIdentifier(const IDRipper::IdentificationRuns& id_runs, 
          const PeptideIdentification& pep_id, 
          const std::map<String, UInt>& file_origin_map, 
          const IDRipper::OriginAnnotationFormat origin_annotation_fmt, 
          bool split_ident_runs);
 
        UInt getIdentRunIdx() const;
 
        UInt getFileOriginIdx() const;
 
        const String & getOriginFullname() const;
 
        const String & getOutputBasename() const;
    };
 
    struct RipFileIdentifierIdxComparator
    {
        bool operator()(const RipFileIdentifier& left, const RipFileIdentifier& right) const;
    };
 
    struct OPENMS_DLLAPI RipFileContent
    {
        std::vector<ProteinIdentification> prot_idents;
        std::vector<PeptideIdentification> pep_idents;
        RipFileContent(const std::vector<ProteinIdentification>& prot_idents, const std::vector<PeptideIdentification>& pep_idents)
            : prot_idents(prot_idents), pep_idents(pep_idents) {}
        const std::vector<ProteinIdentification> & getProteinIdentifications();
        const std::vector<PeptideIdentification> & getPeptideIdentifications();
    };
 
    typedef std::map<RipFileIdentifier, RipFileContent, RipFileIdentifierIdxComparator> RipFileMap;
 
    IDRipper();
 
    ~IDRipper() override;
 
    void rip(
            RipFileMap& ripped,
            std::vector<ProteinIdentification>& proteins,
            std::vector<PeptideIdentification>& peptides,
            bool numeric_filenames,
            bool split_ident_runs);
 
    // Autowrap compatible wrapper for rip(RipFileMap,...)
    void rip(
            std::vector<RipFileIdentifier> & rfis,
            std::vector<RipFileContent> & rfcs,
            std::vector<ProteinIdentification>& proteins,
            std::vector<PeptideIdentification>& peptides,
            bool numeric_filenames,
            bool split_ident_runs);
 
private:
    // Not implemented
    IDRipper(const IDRipper & rhs);
 
    // Not implemented
    IDRipper & operator=(const IDRipper & rhs);
 
    OriginAnnotationFormat detectOriginAnnotationFormat_(std::map<String, UInt> & file_origin_map, const std::vector<PeptideIdentification> & peptide_idents);
    void getProteinHits_(std::vector<ProteinHit> & result, const std::unordered_map<String, const ProteinHit*> & acc2protein_hits, const std::set<String> & protein_accessions);
    std::set<String> getProteinAccessions_(const std::vector<PeptideHit> & peptide_hits);
    int getProteinIdentification_(const PeptideIdentification& pep_ident, const IdentificationRuns& id_runs);
    bool registerBasename_(std::map<String, std::pair<UInt, UInt> >& basename_to_numeric, const IDRipper::RipFileIdentifier& rfi);
    bool setOriginAnnotationMode_(short& mode, short const new_value);
  };
 
} // namespace OpenMS