StringFileFeatures.h

Go to the documentation of this file.

/*
 * This program is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License as published by
 * the Free Software Foundation; either version 3 of the License, or
 * (at your option) any later version.
 *
 * Written (W) 2009 Soeren Sonnenburg
 * Copyright (C) 2009 Berlin Institute of Technology
 */

#ifndef _CSTRINGFILEFEATURES__H__
#define _CSTRINGFILEFEATURES__H__

#include <shogun/lib/config.h>

#include <shogun/features/StringFeatures.h>
#include <shogun/features/Alphabet.h>
#include <shogun/io/MemoryMappedFile.h>
#include <shogun/mathematics/Math.h>
#include <shogun/io/SGIO.h>

namespace shogun
{
class CAlphabet;
template <class T> class CMemoryMappedFile;

template <class ST> class CStringFileFeatures : public CStringFeatures<ST>
{
    public:

    CStringFileFeatures();

    CStringFileFeatures(const char* fname, EAlphabet alpha);

    virtual ~CStringFileFeatures();

    virtual const char* get_name() const { return "StringFileFeatures"; }

    protected:
    ST* get_line(uint64_t& len, uint64_t& offs, int32_t& line_nr, uint64_t file_length);

    virtual void cleanup();

    virtual void cleanup_feature_vector(int32_t num);

    void fetch_meta_info_from_file(int32_t granularity=1048576);

    protected:
    CMemoryMappedFile<ST>* file;
};
}
#endif // _CSTRINGFILEFEATURES__H__