canvasgraphengine
v1.0.4
Published
A library to create graphs similar to Unreal Blueprints.
Downloads
6
Readme
CanvasGraphEngine
A library to create graphs similar to Unreal Blueprints.
Inspired by litegraph.js.
Load data, train, predict and layer feature extraction of a ML MNIST classifier using CanvasGraphEngine.
Features
- Uses Canvas for optimized performance.
- Custom widgets.
- Custom node input/output types.
- Touch support.
- Serialization/Deserialization.
- Subgraphs support.
- Fully typed.
Installation
Node
yarn add canvasgraphengineBrowser
<script src="https://cdn.jsdelivr.net/gh/AIFAnatic/canvasgraphengine@latest/dist/canvasgraphengine.js"></script>Usage
Browser
<html>
<head>
<style>
html, body {
width: 100%;
height: 100%;
margin: 0;
}
#graphcanvas {
background-color: #202020;
}
</style>
</head>
<body>
<div style="width: 100%; height: 100%;">
<canvas id="graphcanvas"></canvas>
</div>
<script type="module">
import {Graph, GraphNode} from '../dist/esm/canvasgraphengine-esm.js'
const canvas = document.getElementById("graphcanvas");
const graph = new Graph(canvas);
graph.registerNode("test", GraphNode);
const node1 = graph.createNode("test", "Node 1");
const node2 = graph.createNode("test", "Node 2");
const node3 = graph.createNode("test", "Node 3");
node2.properties.position.x = 250;
node2.properties.position.y = 250;
node3.properties.position.x = 350;
const node3Input = node3.addInput("input", "");
const node3Output = node3.addOutput("output", "");
const node1Output = node1.addOutput("output", "");
const node1Output1 = node1.addOutput("output", "type");
const node2Input = node2.addInput("input", "");
const node2Input1 = node2.addInput("input1", "type");
const node1Node2connection = node1.connect(node1Output, node2Input);
const node1Node2connection2 = node1.connect(node1Output, node3Input);
setInterval(() => {
node1.triggerOutput(0);
}, 3000);
</script>
</body>
</html>Module
import {Graph, GraphNode} from 'canvasgraphengine';
const canvas = document.getElementById("graphcanvas");
canvas.width = document.body.clientWidth;
canvas.height = document.body.clientHeight;
const graph = new Graph(canvas);
graph.registerNode("test", GraphNode);
const node = graph.createNode("test", "Sample node");Custom node
import {Graph, GraphNode, GraphNodeOutput, GraphNodeInput} from 'canvasgraphengine';
class SampleNode extends GraphNode {
constructor(graph, path) {
super(graph, path, "SampleNode");
this.addInput("input", "text");
this.addOutput("output", "text");
}
public onTrigger() {
console.log("Node was triggered");
}
public onConnectionsChange(from: GraphNodeOutput, to: GraphNodeInput) {
this.triggerOutput(0);
}
}
const canvas = document.getElementById("graphcanvas");
canvas.width = document.body.clientWidth;
canvas.height = document.body.clientHeight;
const graph = new Graph(canvas);
graph.registerNode("test", SampleNode);
const node = graph.createNode("test");Custom widget
import {GraphNode, INodeWidget} from 'canvasgraphengine';
export class TextWidget implements INodeWidget {
public properties: INodeWidgetProperties;
private node: GraphNode;
constructor(node: GraphNode) {
this.properties = NodeWidgetProperties.default();
this.node = node;
}
public draw(ctx: CanvasRenderingContext2D) {
ctx.save();
ctx.fillRect(this.properties.position.x, this.properties.position.y, this.properties.size.w, this.properties.size.h);
}
}Serialization/Deserialization
import {Graph} from 'canvasgraphengine';
const canvas = document.getElementById("graphcanvas");
const graph = new Graph(canvas);
const json = graph.toJSON(); // Serialize
graph.fromJSON(json) // DeserializeTODO
- [ ] Docs