Open
Description
Hi,
After viewing the Deno 2.0 anoncement video, I tryed to use observablehq's samples in deno Jupyter Notebooks.
some samplkes use canvas, and so do crash.
by implementing some ugly code I can make canvas's dependent code to works, so it's posible to use an EmulatedCanvas2D to make canvas works in linkdom.
the question is: Can a canvas like can be integrated to the project ?
for reference there is a deno Notebooks of plot-mandelbrot-set:
import * as Plot from "npm:@observablehq/plot";
import { document } from "jsr:@ry/jupyter-helper";
import { createCanvas, type EmulatedCanvas2D } from "https://deno.land/x/canvas@v1.4.2/mod.ts";
function mandelbrot(x: number, y: number) {
for (let n = 0, zr = 0, zi = 0; n < 80; ++n) {
[zr, zi] = [zr * zr - zi * zi + x, 2 * zr * zi + y];
if (zr * zr + zi * zi > 4) return n;
}
}
const old = document.createElement;
document.createElement = (...args: any[]) => {
if (args.length === 1 && args[0] === "canvas") {
let realCanvas: EmulatedCanvas2D | undefined;
const canvas = {
height: 200,
width: 200,
getContext: (drv: string) => {
realCanvas = createCanvas(canvas.width, canvas.height);
return realCanvas.getContext(drv);
},
toDataURL: () => {
return realCanvas?.toDataURL();
}
}
return canvas;
}
return old.apply(document, args);
}
Plot.raster({fill: mandelbrot, x1: -2, x2: 1, y1: -1.164, y2: 1.164}).plot({aspectRatio: 1, document})Metadata
Metadata
Assignees
Labels
No labels