InspectOutfile.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: Martin Langwisch $
// --------------------------------------------------------------------------
 
#pragma once
 
#include <OpenMS/CONCEPT/Exception.h>
#include <OpenMS/DATASTRUCTURES/String.h>
#include <OpenMS/KERNEL/MSExperiment.h>
#include <OpenMS/KERNEL/StandardTypes.h>
#include <OpenMS/METADATA/PeptideIdentification.h>
#include <OpenMS/FORMAT/FileHandler.h>
#include <OpenMS/FORMAT/FileTypes.h>
 
 
namespace OpenMS
{
  class OPENMS_DLLAPI InspectOutfile
  {
public:
    InspectOutfile();
 
    InspectOutfile(const InspectOutfile & inspect_outfile);
 
    virtual ~InspectOutfile();
 
    InspectOutfile & operator=(const InspectOutfile & inspect_outfile);
 
    bool operator==(const InspectOutfile & inspect_outfile) const;
 
    std::vector<Size> load(const String & result_filename, std::vector<PeptideIdentification> & peptide_identifications, ProteinIdentification & protein_identification, const double p_value_threshold, const String & database_filename = "");
 
    std::vector<Size> getWantedRecords(const String & result_filename, double p_value_threshold);
 
    void compressTrieDB(const String & database_filename, const String & index_filename, std::vector<Size> & wanted_records, const String & snd_database_filename, const String & snd_index_filename, bool append = false);
 
    void generateTrieDB(const String & source_database_filename, const String & database_filename, const String & index_filename, bool append = false, const String& species = "");
 
 
    void getACAndACType(String line, String & accession, String & accession_type);
 
    void getPrecursorRTandMZ(const std::vector<std::pair<String, std::vector<std::pair<Size, Size> > > > & files_and_peptide_identification_with_scan_number, std::vector<PeptideIdentification> & ids);
 
    void getLabels(const String & source_database_filename, String & ac_label, String & sequence_start_label, String & sequence_end_label, String & comment_label, String & species_label);
 
    std::vector<Size> getSequences(const String & database_filename, const std::map<Size, Size> & wanted_records, std::vector<String> & sequences);
 
    void getExperiment(PeakMap & exp, String & type, const String & in_filename)
    {
      type.clear();
      exp.reset();
      //input file type
      FileHandler fh;
      FileTypes::Type in_type = fh.getTypeByContent(in_filename);
      if (in_type == FileTypes::UNKNOWN)
      {
        throw Exception::ParseError(__FILE__, __LINE__, OPENMS_PRETTY_FUNCTION, "Could not determine type of the file. Aborting!", in_filename);
      }
      type = FileTypes::typeToName(in_type);
      fh.loadExperiment(in_filename, exp, in_type, ProgressLogger::NONE, false, false);
    }
 
    bool getSearchEngineAndVersion(const String & cmd_output, ProteinIdentification & protein_identification);
 
    void readOutHeader(const String & filename, const String & header_line, Int & spectrum_file_column, Int & scan_column, Int & peptide_column, Int & protein_column, Int & charge_column, Int & MQ_score_column, Int & p_value_column, Int & record_number_column, Int & DB_file_pos_column, Int & spec_file_pos_column, Size & number_of_columns);
 
protected:
    static const Size db_pos_length_;         
    static const Size trie_db_pos_length_;         
    static const Size protein_name_length_;         
    static const Size record_length_;         
    static const char trie_delimiter_;         
    static const String score_type_;        
  };
 
} //namespace OpenMS