node package manager


VivaGraph Build Status

VivaGraphJS is designed to be extensible and to support different rendering engines and layout algorithms. Underlying algorithms have been broken out into ngraph.

The larger family of modules can be found by querying npm for "ngraph".


Enough talking. Show me the demo!

Some examples of library usage in the real projects:

To start using the library include vivagraph.js script from the dist folder. The following code is the minimum required to render a graph with two nodes and one edge:

var graph = Viva.Graph.graph();
graph.addLink(1, 2);
var renderer = Viva.Graph.View.renderer(graph);;

This will instantiate a graph inside document.body:

Simple graph

If you want to render graph in your own DOM element:

var graph = Viva.Graph.graph();
graph.addLink(1, 2);
// specify where it should be rendered: 
var renderer = Viva.Graph.View.renderer(graph, {
  container: document.getElementById('graphDiv')

The code above adds a link to the graph between nodes 1 and 2. Since nodes are not yet in the graph they will be created. It's equivalent to

var graph = Viva.Graph.graph();
graph.addLink(1, 2);
var renderer = Viva.Graph.View.renderer(graph);;


VivaGraphJS is all about customization. You can easily change the appearance of nodes and links. You can also change the layouting algorithm and medium that displays elements of the graph. For example: The following code allows you to use WebGL-based rendering, instead of the default SVG.

var graph = Viva.Graph.graph();
graph.addLink(1, 2);
var graphics = Viva.Graph.View.webglGraphics();
var renderer = Viva.Graph.View.renderer(graph,
        graphics : graphics

graphics class is responsible for rendering nodes and links on the page. And renderer orchestrates the process. To change nodes appearance tell graphics how to represent them. Here is an example of graph with six people who I follow at github:

var graph = Viva.Graph.graph();
// Construct the graph 
graph.addNode('anvaka', {url : ''});
graph.addNode('manunt', {url : ''});
graph.addNode('thlorenz', {url : ''});
graph.addNode('bling', {url : ''});
graph.addNode('diyan', {url : ''});
graph.addNode('pocheptsov', {url : ''});
graph.addNode('dimapasko', {url : ''});
graph.addLink('anvaka', 'manunt');
graph.addLink('anvaka', 'thlorenz');
graph.addLink('anvaka', 'bling');
graph.addLink('anvaka', 'diyan');
graph.addLink('anvaka', 'pocheptsov');
graph.addLink('anvaka', 'dimapasko');
// Set custom nodes appearance 
var graphics = Viva.Graph.View.svgGraphics();
graphics.node(function(node) {
       // The function is called every time renderer needs a ui to display node 
       return Viva.Graph.svg('image')
             .attr('width', 24)
             .attr('height', 24)
             .link(; // holds custom object passed to graph.addNode(); 
    .placeNode(function(nodeUI, pos){
        // Shift image to let links go to the center: 
        nodeUI.attr('x', pos.x - 12).attr('y', pos.y - 12);
var renderer = Viva.Graph.View.renderer(graph,
        graphics : graphics

The result is:

Custom nodes

Tuning layout algorithm

Graphs vary by their nature. Some graphs have hundreds of nodes and few edges (or links), some might connect every node with every other. Tuning the physics often helps get the best layout. Consider the following example:

var graphGenerator = Viva.Graph.generator();
var graph = graphGenerator.grid(3, 3);
var renderer = Viva.Graph.View.renderer(graph);;

Graph generators are part of the library, which can produce classic graphs. grid generator creates a grid with given number of columns and rows. But with default parameters the rendering is pretty ugly:

Grid 3x3 bad

Let's tweak the original code:

var graphGenerator = Viva.Graph.generator();
var graph = graphGenerator.grid(3, 3);
var layout = Viva.Graph.Layout.forceDirected(graph, {
    springLength : 10,
    springCoeff : 0.0005,
    dragCoeff : 0.02,
    gravity : -1.2
var renderer = Viva.Graph.View.renderer(graph, {
    layout : layout

Now the result is much better:

Grid 3x3

You can tune values during simulation with layout.simulator.springLength(newValue), layout.simulator.springCoeff(newValue), etc. See all the values that you can tune in this source file.

Tuning layout algorithm is definitely one of the hardest part of using this library. It has to be improved in future to simplify usage. Each of the force directed algorithm parameters are described in the source code.

Design philosophy/roadmap

Until version 0.7.x VivaGraph was a single monolithic code base. Starting from 0.7.x the library is bundled from small npm modules into Viva namespace. All these modules are part of a larger ngraph family. ngraph modules support rendering graphs into images, 3D rendering, integration with gephi, pagerank calculation and many more.

Version 0.7 is a compromise between maximum backward compatibility and ngraph flexibility. Eventually I hope to further simplify API and provide interface for custom builds.

Upgrade guide

Please refer the upgrade guide to see how to update older versions of the library to the latest one.

Local Build

Run the following script:

git clone
cd ./VivaGraphJS
npm install
gulp release

The combined/minified code should be stored in dist folder.

Looking for alternatives?

I'm trying to put up a list of all known graph drawing libraries. Please find it here

I need your feedback

Disclaimer: I wrote this library to learn JavaScript. By no means I pretend to be an expert in the language and chosen approach to design may not be the optimal. I would love to hear your feedback and suggestions.

Though I implemented this library from scratch, I went through many existing libraries to pick the best (at my view) out of them. If you are evaluating libraries for your project make sure to check them out as well.

My goal is to create highly performant javascript library, which serves in the field of graph drawing. To certain extent I achieved it. But I have no doubt there is much more to improve here.