Fasta.h

Go to the documentation of this file.

// SPDX-FileCopyrightText: The Bio++ Development Group
//
// SPDX-License-Identifier: CECILL-2.1
 
#ifndef BPP_SEQ_IO_FASTA_H
#define BPP_SEQ_IO_FASTA_H
 
 
#include "../Container/SequenceContainer.h"
#include "../Container/VectorSequenceContainer.h"
#include "../Sequence.h"
#include "AbstractIAlignment.h"
#include "AbstractISequence.h"
#include "AbstractOSequence.h"
#include "ISequenceStream.h"
#include "OSequenceStream.h"
#include "SequenceFileIndex.h"
 
namespace bpp
{
class Fasta :
  public AbstractISequence,
  public AbstractIAlignment,
  public AbstractOSequence2,
  public virtual ISequenceStream,
  public virtual OSequenceStream
{
protected:
  unsigned int charsByLine_; // Number of char by line (output only)
  bool checkNames_;          // If names must be checked in container
  bool extended_;            // If using HUPO-PSI extensions
  bool strictNames_;         // If name is between '>' and first space
 
public:
  Fasta(unsigned int charsByLine = 100, bool checkSequenceNames = true, bool extended = false, bool strictSequenceNames = false) : charsByLine_(charsByLine), checkNames_(checkSequenceNames), extended_(extended), strictNames_(strictSequenceNames) {}
 
  // Class destructor
  virtual ~Fasta() {}
 
public:
  void appendSequencesFromStream(std::istream& input, SequenceContainerInterface& sc) const override;
  void appendAlignmentFromStream(std::istream& input, SequenceContainerInterface& sc) const override
  {
    appendSequencesFromStream(input, sc); // This may raise an exception if sequences are not aligned!
  }
  void writeSequences(std::ostream& output, const SequenceContainerInterface& sc) const override;
 
  void writeSequences(const std::string& path, const SequenceContainerInterface& sc, bool overwrite = true) const override
  {
    AbstractOSequence::writeSequences(path, sc, overwrite);
  }
  const std::string getFormatName() const override { return "FASTA file"; }
 
  const std::string getFormatDescription() const override
  {
    return "Sequence name (preceded by >) in one line, sequence content, no comments";
  }
  bool nextSequence(std::istream& input, Sequence& seq) const override;
  void writeSequence(std::ostream& output, const Sequence& seq) const override;
  bool checkNames() const { return checkNames_; }
 
  void checkNames(bool yn) { checkNames_ = yn; }
 
  bool strictNames() const { return strictNames_; }
 
  void strictNames(bool yn) { strictNames_ = yn; }
 
  class FileIndex : SequenceFileIndex
  {
public:
    FileIndex() : index_(), fileSize_(0) {}
    ~FileIndex() {}
    void build(const std::string& path)
    {
      build(path, false);
    }
    void build(const std::string& path, const bool strictSequenceNames);
    std::streampos getSequencePosition(const std::string& id) const;
    size_t getNumberOfSequences() const
    {
      return index_.size();
    }
    void read(const std::string& path);
    void write(const std::string& path);
    void getSequence(const std::string& seqid, Sequence& seq, const std::string& path) const;
    void getSequence(const std::string& seqid, Sequence& seq, const std::string& path, const bool strictSequenceNames) const;
 
private:
    std::map<std::string, std::streampos> index_;
    std::streampos fileSize_;
  };
};
} // end of namespace bpp.
#endif // BPP_SEQ_IO_FASTA_H