DBInitializerFromCSV.h

Go to the documentation of this file.

#ifndef GUM_LEARNING_DB_INITILIALIZER_FROM_CSV_H
#define GUM_LEARNING_DB_INITILIALIZER_FROM_CSV_H

#include <vector>
#include <string>
#include <fstream>
#include <iostream>
#include <sstream>

#include <agrum/agrum.h>
#include <agrum/learning/database/IDBInitializer.h>
#include <agrum/learning/database/CSVParser.h>

namespace gum {

  namespace learning {

    template < template < typename > class ALLOC = std::allocator >
    class DBInitializerFromCSV : public IDBInitializer< ALLOC > {
      public:
      using allocator_type = ALLOC< std::string >;

      // ##########################################################################
      // ##########################################################################


      DBInitializerFromCSV(const std::string     filename,
                           bool                  fileContainsNames = true,
                           const std::string     delimiter = ",",
                           const char            commentmarker = '#',
                           const char            quoteMarker = '"',
                           const allocator_type& alloc = allocator_type());


      DBInitializerFromCSV(const DBInitializerFromCSV< ALLOC >& from);


      DBInitializerFromCSV(const DBInitializerFromCSV< ALLOC >& from,
                           const allocator_type&                alloc);

      DBInitializerFromCSV(DBInitializerFromCSV< ALLOC >&& from);

      DBInitializerFromCSV(DBInitializerFromCSV< ALLOC >&& from,
                           const allocator_type&           alloc);

      virtual DBInitializerFromCSV< ALLOC >* clone() const;

      virtual DBInitializerFromCSV< ALLOC >*
         clone(const allocator_type& alloc) const;

      virtual ~DBInitializerFromCSV();



      // ##########################################################################
      // ##########################################################################



      DBInitializerFromCSV< ALLOC >&
         operator=(const DBInitializerFromCSV< ALLOC >& from);


      DBInitializerFromCSV< ALLOC >&
         operator=(DBInitializerFromCSV< ALLOC >&& from);



      protected:
      virtual std::vector< std::string, ALLOC< std::string > >
         _variableNames() final;

      virtual const std::vector< std::string, ALLOC< std::string > >&
         _currentStringRow() final;

      virtual bool _nextRow() final;


#ifndef DOXYGEN_SHOULD_SKIP_THIS

      private:
      // the filename used for parsing
      std::string __filename;

      // indicates the delimiter used within the CSV
      std::string __delimiter;

      // indicates which character is a comment symbol in the CSV
      char __comment_marker;

      // indicates which character is a quote symbol in the CSV
      char __quote_marker;

      // indicates whether the first row of the file contains the names
      bool __first_row_has_names;

      // the input stream read by the parser
      std::ifstream __input_stream;

      // the CSV parser used for the reading the CSV file
      CSVParser< ALLOC > __parser;

      // the variables names, if the first row has names
      std::vector< std::string, ALLOC< std::string > > __var_names;

#endif /* DOXYGEN_SHOULD_SKIP_THIS */
    };

  } /* namespace learning */

} /* namespace gum */

// always include the template implementation
#include <agrum/learning/database/DBInitializerFromCSV_tpl.h>


#endif /* GUM_LEARNING_DB_INITILIALIZER_FROM_CSV_H */