Giraffe 🦒

Experimental graphing library in go

a list of what is implemented so far :-

• Search

  • Breadth-First
  • Depth-First

• Sorting (Siblings, Nodes)

• Centrality

  • Degree
  • Betweenness

• Clustering

  • K-Means

• Exporters

  • Export to Mermaid
  • Export to Graphviz Dot

• What is a graphing library?

• Examples

  • Depth-first search
  • Get various centralities for the graph
  • K-Means Clustering
  • Export to various formats

Note

I haven't used this library for serious experimentations with actual graph data. A lot of this is purely through intuition, eg - this is how it's supposed to work if I understand it correctly.

Usage

Simplest representation of a graph

package main 

import (
    "github.com/aadv1k/giraffe"
    "fmt"
)

func main() {
    var g giraffe.Graph

    g.AddVertex(&giraffe.Vertex{ Index: 0 })
    g.AddVertex(&giraffe.Vertex{ Index: 1 })
    g.AddVertex(&giraffe.Vertex{ Index: 2 })

    g.AddEdge(&giraffe.Edge{Start: g.GetVertex(0), End: g.GetVertex(1)})
    g.AddEdge(&giraffe.Edge{Start: g.GetVertex(1), End: g.GetVertex(2)})
    g.AddEdge(&giraffe.Edge{Start: g.GetVertex(2), End: g.GetVertex(0)})
    
    fmt.Print(g.ParseToMermaid())
}

graph TD;
    0;
    1;
    2;
    0 --> 1;
    1 --> 2;
    2 --> 0;

Loading

What is a graphing library?

Graphing library implement a host of functions that interact with a data structure called a graph. A graph consists of Vertices and Edges, the latter defines how one Vertex relates to other.

Mature libraries such as NetworkX, often deal with graphs spanning over a million vertices with amazing efficiency and speed.

This is not such library. It is primarily built as an exploratory/research project to probe this paradign of computation

Examples

A Depth-first search

package main

import (
	"github.com/aadv1k/giraffe"
)

func main() {
	var g *giraffe.Graph
	g = giraffe.MakePyramidGraph() // Utility function to generate a sample graph

	// Visit the vertices using Depth-First Search
	_, visited := g.FindVertexDFS(3)

	giraffe.PrintVertices(visited) // Utility function to print the vertex array
}

{ 0, 2, 5, 4, 3 }

Additionally, you can also do the same using Breadth-First Search

    // ... 

    _, visited := g.FindVertexBFS(0, 3)

    // ...

{ 0, 1, 2, 3 }

Get various centralities for the graph

func main() {
	var g *giraffe.Graph
	g = giraffe.MakePyramidGraph() // Utility function to generate a sample graph

	fmt.Printf("Degree: %v\n", g.GetDegree())
	fmt.Printf("Betweenness: %v\n", g.GetBetweenness())
}

Degree: [0 1 1 1 1 2]
Betweenness: [1 1 1 1 1 1]

K-Means Clustering

func main() {
	var g *giraffe.Graph
	g = giraffe.MakeClusterGraph() // Utility function to generate a sample graph

	clusters, means := g.KMeansClustering(2)

	fmt.Printf("%v\n", means)
	fmt.Printf("There are %d clusters", len(clusters))
}

[1.472636459754717 1.0380631836067153]
There are 2 clusters

Export to various formats

The library doesn't come with any dependency, instead it can spit out DSL code that can allow you to render the graph in other libs. Eg

func main() {
	var g *giraffe.Graph
	g = giraffe.MakeClusterGraph()

	fmt.Print(g.ParseToMermaid())
}

graph TD;
    0;
    1;
    2;
    3;
    4;
    5;
    0 --> 1;
    1 --> 2;
    2 --> 0;
    3 --> 0;
    3 --> 4;
    3 --> 5;
    5 --> 4;

Loading

or Graphviz

func main() {
	var g *giraffe.Graph
	g = giraffe.MakeClusterGraph()

	fmt.Print(g.ParseToGraphvizDot())
}

digraph G {
    0;
    1;
    2;
    3;
    4;
    5;
    0 -> 1;
    1 -> 2;
    2 -> 0;
    3 -> 0;
    3 -> 4;
    3 -> 5;
    5 -> 4;
}

and that's it for now! these functions describe the upper and lower-bound of this experimental library.