Skip to main content

How to resize and rotate canvas shapes with React and Konva?

Currently there is no good, pure declarative "react-way" to use the Transformer tool. But you can still use it with some small manual requests to the Konva nodes. And it will work just fine.

The idea: you need to create a Konva.Transformer node, and attach it to the required node manually.

Instructions: Click on one of the rectangles to select it. Then you can:

  • Drag it to move
  • Use handles to resize
  • Click outside to deselect
import React from 'react';
import { Stage, Layer, Rect, Transformer } from 'react-konva';

const Rectangle = ({ shapeProps, isSelected, onSelect, onChange }) => {
const shapeRef = React.useRef();
const trRef = React.useRef();

React.useEffect(() => {
if (isSelected) {
// we need to attach transformer manually
trRef.current.nodes([shapeRef.current]);
}
}, [isSelected]);

return (
<React.Fragment>
<Rect
onClick={onSelect}
onTap={onSelect}
ref={shapeRef}
{...shapeProps}
draggable
onDragEnd={(e) => {
onChange({
...shapeProps,
x: e.target.x(),
y: e.target.y(),
});
}}
onTransformEnd={(e) => {
// transformer is changing scale of the node
// and NOT its width or height
// but in the store we have only width and height
// to match the data better we will reset scale on transform end
const node = shapeRef.current;
const scaleX = node.scaleX();
const scaleY = node.scaleY();

// we will reset it back
node.scaleX(1);
node.scaleY(1);
onChange({
...shapeProps,
x: node.x(),
y: node.y(),
// set minimal value
width: Math.max(5, node.width() * scaleX),
height: Math.max(5, node.height() * scaleY),
});
}}
/>
{isSelected && (
<Transformer
ref={trRef}
flipEnabled={false}
boundBoxFunc={(oldBox, newBox) => {
// limit resize
if (Math.abs(newBox.width) < 5 || Math.abs(newBox.height) < 5) {
return oldBox;
}
return newBox;
}}
/>
)}
</React.Fragment>
);
};

const initialRectangles = [
{
x: 10,
y: 10,
width: 100,
height: 100,
fill: 'red',
id: 'rect1',
},
{
x: 150,
y: 150,
width: 100,
height: 100,
fill: 'green',
id: 'rect2',
},
];

const App = () => {
const [rectangles, setRectangles] = React.useState(initialRectangles);
const [selectedId, selectShape] = React.useState(null);

const checkDeselect = (e) => {
// deselect when clicked on empty area
const clickedOnEmpty = e.target === e.target.getStage();
if (clickedOnEmpty) {
selectShape(null);
}
};

return (
<Stage
width={window.innerWidth}
height={window.innerHeight}
onMouseDown={checkDeselect}
onTouchStart={checkDeselect}
>
<Layer>
{rectangles.map((rect, i) => {
return (
<Rectangle
key={i}
shapeProps={rect}
isSelected={rect.id === selectedId}
onSelect={() => {
selectShape(rect.id);
}}
onChange={(newAttrs) => {
const rects = rectangles.slice();
rects[i] = newAttrs;
setRectangles(rects);
}}
/>
);
})}
</Layer>
</Stage>
);
};

export default App;

What Transformer does not do

Transformer draws the handles and applies the scale. Snapping to other objects, alignment guides, a shared bounding box for a multi-selection, and per-shape aspect rules are all yours to build — see objects snapping for one approach.