StreamingAsciiFile.h

浏览该文件的文档.

/*
 * 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) 2011 Shashwat Lal Das
 * Copyright (C) 2011 Berlin Institute of Technology and Max-Planck-Society
 */
#ifndef __STREAMING_ASCIIFILE_H__
#define __STREAMING_ASCIIFILE_H__

#include <shogun/lib/config.h>

#include <shogun/io/streaming/StreamingFile.h>
#include <shogun/lib/v_array.h>

namespace shogun
{

struct substring;
template <class ST> struct SGSparseVectorEntry;
template <class T> class DynArray;

class CStreamingAsciiFile: public CStreamingFile
{

public:
    CStreamingAsciiFile();

    CStreamingAsciiFile(const char* fname, char rw='r');

    virtual ~CStreamingAsciiFile();

    void set_delimiter(char delimiter);

#ifndef SWIG // SWIG should skip this

    inline bool str_to_bool(char *str)
    {
        return (atoi(str)!=0);
    }

#define GET_VECTOR_DECL(sg_type)                    \
    virtual void get_vector                     \
        (sg_type*& vector, int32_t& len);           \
                                    \
    virtual void get_vector_and_label               \
        (sg_type*& vector, int32_t& len, float64_t& label); \
                                    \
    virtual void get_string                     \
        (sg_type*& vector, int32_t& len);           \
                                    \
    virtual void get_string_and_label               \
        (sg_type*& vector, int32_t& len, float64_t& label); \
                                    \
    virtual void get_sparse_vector                  \
        (SGSparseVectorEntry<sg_type>*& vector, int32_t& len);  \
                                    \
    virtual void get_sparse_vector_and_label            \
        (SGSparseVectorEntry<sg_type>*& vector, int32_t& len, float64_t& label);

    GET_VECTOR_DECL(bool)
    GET_VECTOR_DECL(uint8_t)
    GET_VECTOR_DECL(char)
    GET_VECTOR_DECL(int32_t)
    GET_VECTOR_DECL(float32_t)
    GET_VECTOR_DECL(float64_t)
    GET_VECTOR_DECL(int16_t)
    GET_VECTOR_DECL(uint16_t)
    GET_VECTOR_DECL(int8_t)
    GET_VECTOR_DECL(uint32_t)
    GET_VECTOR_DECL(int64_t)
    GET_VECTOR_DECL(uint64_t)
    GET_VECTOR_DECL(floatmax_t)
#undef GET_VECTOR_DECL

#endif // #ifndef SWIG // SWIG should skip this

    virtual const char* get_name() const
    {
        return "StreamingAsciiFile";

    }

private:
    template <class T> void append_item(DynArray<T>* items, char* ptr_data, char* ptr_item);

    void tokenize(char delim, substring s, v_array<substring> &ret);

private:
    v_array<substring> words;

    char m_delimiter;
};
}
#endif //__STREAMING_ASCIIFILE_H__