vector.h

//
//  vector.h
//  Perceptron GLM NLP Tasks
//
//  Created by husnu sensoy on 19/02/14.
//  Copyright (c) 2014 husnu sensoy. All rights reserved.
//

#ifndef Perceptron_GLM_NLP_Tasks_vector_h
#define Perceptron_GLM_NLP_Tasks_vector_h

#include <float.h>
#define NEGATIVE_INFINITY (-(FLT_MAX - 10.))

#include <stdlib.h>
#include <stdbool.h>
#include <math.h>

#include "uthash.h"
#include "debug.h"

struct vector {
    size_t n;
    size_t last_idx;
    float* data;
};

typedef struct vector* vector;

vector parse_vector(char *buff);

vector vector_create(size_t n);
void vector_free(vector);



/**
 * @brief Element-wise vector addition: target = target + mult * src
 * @param target Target vector on which src to be added.
 * @param src Vector to be added.
 * @param mult Multiplier of src before adding on target
 */
void vadd(vector target, const vector src, float mult);

float vdot(vector v1, vector v2);

void vdiv(vector v, int div);
void vnorm(vector v, size_t n);

void vprint(vector v);

vector vlinear(vector target, vector src);
vector vquadratic(vector target, vector src, float d);
vector vcubic(vector target, vector src, size_t length);

/**
 * @brief Vector concatenation like np.concatenate
 * @param target 
 * @param v
 * @return vector of true_n = target->true_n + v->true_n aligned to 64 bytes
 */
vector vconcat(vector target, const vector v);

vector vconcat_arr(vector target, size_t n, const float arr[]);

/**
 * 
 * @param v1
 * @param v2
 * @return v1.v2
 */
static inline float linear(vector v1, vector v2) {
    return vdot(v1, v2);
}

/**
 * 
 * @param v1
 * @param v2
 * @param d 
 * @param n
 * @return (v1.v2 + d )^n
 */
static inline float polynomial(vector v1, vector v2, float d, int n) {
    return pow(vdot(v1, v2) + d, n);
}
#endif