FASTAFile.h

Go to the documentation of this file.

// Copyright (c) 2002-present, OpenMS Inc. -- EKU Tuebingen, ETH Zurich, and FU Berlin
// SPDX-License-Identifier: BSD-3-Clause
//
// --------------------------------------------------------------------------
// $Maintainer: Chris Bielow $
// $Authors: Chris Bielow, Nora Wild $
// --------------------------------------------------------------------------
 
#pragma once
 
#include <OpenMS/CONCEPT/Exception.h>
#include <OpenMS/DATASTRUCTURES/StringUtils.h>
#include <OpenMS/CONCEPT/ProgressLogger.h>
 
#include <fstream>
#include <utility>
#include <vector>
 
namespace OpenMS
{
    class OPENMS_DLLAPI FASTAFile : public ProgressLogger
    {
    public:
        struct FASTAEntry
        {
            std::string identifier;
            std::string description;
            std::string sequence;
 
            FASTAEntry() = default;
 
            FASTAEntry(const std::string& id, const std::string& desc, const std::string& seq) :
                    identifier(id),
                    description(desc),
                    sequence(seq)
            {
            }
 
            FASTAEntry(const FASTAEntry& rhs) = default;
 
            FASTAEntry(FASTAEntry&& rhs) noexcept 
                    :
                    identifier(::std::move(rhs.identifier)),
                    description(::std::move(rhs.description)),
                    sequence(::std::move(rhs.sequence))
            {
            }
 
 
            FASTAEntry& operator=(const FASTAEntry& rhs) = default;
 
            bool operator==(const FASTAEntry& rhs) const
            {
                return identifier == rhs.identifier
                       && description == rhs.description
                       && sequence == rhs.sequence;
            }
 
            bool headerMatches(const FASTAEntry& rhs) const
            {
                return identifier == rhs.identifier &&
                       description == rhs.description;
            }
 
            bool sequenceMatches(const FASTAEntry& rhs) const
            {
                return sequence == rhs.sequence;
            }
        };
 
        FASTAFile() = default;
 
        ~FASTAFile() override = default;
 
        void readStart(const std::string& filename);
 
        void readStartWithProgress(const std::string& filename, const std::string& progress_label);
 
        bool readNext(FASTAEntry& protein);
 
        bool readNextWithProgress(FASTAEntry& protein);
 
        std::streampos position();
 
        bool atEnd();
 
        bool setPosition(const std::streampos& pos);
 
        void writeStart(const std::string& filename);
 
        void writeNext(const FASTAEntry& protein);
 
        void writeEnd();
 
 
        void load(const std::string& filename, std::vector<FASTAEntry>& data) const;
 
        void store(const std::string& filename, const std::vector<FASTAEntry>& data) const;
 
    protected:
        bool readEntry_(std::string& id, std::string& description, std::string& seq);
 
        std::fstream infile_;       
        std::ofstream outfile_;     
        Size entries_read_{0};      
        std::streampos fileSize_{}; 
        std::string seq_;           
        std::string id_;            
        std::string description_;   
    };
 
} // namespace OpenMS