vectorizer

README

Vectorizer

Vectorizer is a generic compatible module, which provides an efficient way to manage and manipulate sparse vectors. It supports operations such as applying values to vectors, calculating magnitudes, and computing cosine similarity.

Features

• Sparse Vector Representation: Efficiently store vectors with non-zero values only.
• Vector Operations: Includes methods for scaling, normalizing, and calculating cosine similarity.

Prerequisites

• Go 1.23.0 or above

Installation

go get -u github.com/softwarespot/vectorizer

Usage

A basic example of using Vectorizer.

package main

import (
	"fmt"

	"github.com/softwarespot/vectorizer"
)

func main() {
	// Create a vectorizer with the type "string"
   	vz := vectorizer.New[string](128)

	vec1 := vectorizer.NewVector[string]()
	vz.ApplyTo(vec1, "Event 1", 1)
	vz.ApplyTo(vec1, "Event 2", 2)
	vz.ApplyTo(vec1, "Event 3", 3)

	vec2 := vectorizer.NewVector[string]()
	vz.ApplyTo(vec2, "Event 2", 2)
	vz.ApplyTo(vec2, "Event 3", 3)

	fmt.Println(vec1.CosineSimilarity(vec2)) // outputs: 0.9636241116594316
}

License

The code has been licensed under the MIT license.

Documentation

Index

• type Vector
  • func NewVector[T comparable]() *Vector[T]
  • func (v *Vector[T]) CosineSimilarity(v2 *Vector[T]) (float64, error)
  • func (v *Vector[T]) Normalize()
  • func (v *Vector[T]) Scale(scalar float64)
• type Vectorizer
  • func New[T comparable](size int) *Vectorizer[T]
  • func (vz *Vectorizer[T]) ApplyTo(v *Vector[T], key T, value float64)

Types

type Vector

type Vector[T comparable] struct {
	// contains filtered or unexported fields
}

Vector represents a sparse vector.

func NewVector

func NewVector[T comparable]() *Vector[T]

NewVector creates a new empty Vector.

func (*Vector[T]) CosineSimilarity

func (v *Vector[T]) CosineSimilarity(v2 *Vector[T]) (float64, error)

CosineSimilarity calculates the cosine similarity between two vectors. It returns an error if either vector has zero magnitude.

func (*Vector[T]) Normalize

func (v *Vector[T]) Normalize()

Normalize scales the vector to have a magnitude of 1. If the vector has zero magnitude, it remains unchanged.

func (*Vector[T]) Scale

func (v *Vector[T]) Scale(scalar float64)

Scale multiplies all elements of the vector by a scalar value.

type Vectorizer

type Vectorizer[T comparable] struct {
	// contains filtered or unexported fields
}

Vectorizer is a structure that manages dimensions for vectors.

func New

func New[T comparable](size int) *Vectorizer[T]

New creates a new Vectorizer with an initial size.

func (*Vectorizer[T]) ApplyTo

func (vz *Vectorizer[T]) ApplyTo(v *Vector[T], key T, value float64)

ApplyTo applies a value to the vector at the specified key.