StreamingHashedDocDotFeatures.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) 2013 Evangelos Anagnostopoulos
 * Copyright (C) 2013 Evangelos Anagnostopoulos
 */
#ifndef _STREAMING_HASHEDDOCDOTFEATURES__H__
#define _STREAMING_HASHEDDOCDOTFEATURES__H__

#include <shogun/features/StringFeatures.h>
#include <shogun/features/streaming/StreamingDotFeatures.h>
#include <shogun/lib/Tokenizer.h>
#include <shogun/converter/HashedDocConverter.h>
#include <shogun/io/streaming/InputParser.h>
#include <shogun/io/streaming/StreamingFileFromStringFeatures.h>

namespace shogun
{
class CStreamingDotFeatures;
class CTokenizer;
class CHashedDocConverter;
 
class CStreamingHashedDocDotFeatures : public CStreamingDotFeatures
{
public:
    CStreamingHashedDocDotFeatures();

    CStreamingHashedDocDotFeatures(CStreamingFile* file, bool is_labelled, int32_t size,
            CTokenizer* tzer, int32_t bits=20);

    CStreamingHashedDocDotFeatures(CStringFeatures<char>* dot_features,CTokenizer* tzer,
            int32_t bits=20, float64_t* lab=NULL);

    virtual ~CStreamingHashedDocDotFeatures();

    virtual float32_t dot(CStreamingDotFeatures* df);

    virtual float32_t dense_dot(const float32_t* vec2, int32_t vec2_len);

    virtual void add_to_dense_vec(float32_t alpha, float32_t* vec2,
            int32_t vec2_len, bool abs_val=false);

    virtual int32_t get_dim_feature_space() const;

    virtual const char* get_name() const;

    virtual int32_t get_num_vectors() const;

    virtual CFeatures* duplicate() const;

    virtual void set_vector_reader();

    virtual void set_vector_and_label_reader();

    virtual EFeatureType get_feature_type() const;

    virtual EFeatureClass get_feature_class() const;

    virtual void start_parser();

    virtual void end_parser();

    virtual float64_t get_label();

    virtual bool get_next_example();

    virtual void release_example();

    virtual int32_t get_num_features();

    SGSparseVector<float64_t> get_vector();

    void set_normalization(bool normalize);

    void set_k_skip_n_grams(int32_t k, int32_t n);

private:
    void init(CStreamingFile* file, bool is_labelled, int32_t size, CTokenizer* tzer,
        int32_t bits, bool normalize, int32_t n_grams, int32_t skips);

protected:
    
    int32_t num_bits;

    SGSparseVector<float64_t> current_vector;

    CTokenizer *tokenizer;

    CHashedDocConverter* converter;

    CInputParser<char> parser;

    float64_t current_label;
};
}

#endif // _STREAMING_HASHEDDOCDOTFEATURES__H__