With today’s code, you can perform a basic validation of generated and loaded NeoBoards. It uses the library Zod which can parse unknown values in TypeScript.
Watch out that this type is slightly incomplete and, additionally, may become outdated when new versions of NeoBoard get released. It’s still a good starting point and able to parse most current NeoBoards.
import z from 'zod';
export const zPoint = z.object({
x: z.number(),
y: z.number(),
});
export const zWhiteboard = z.object({
version: z.literal('net.nordeck.whiteboard@v1'),
whiteboard: z.object({
files: z.array(z.object({
mxc: z.string(),
data: z.string(),
})).optional(),
slides: z.array(z.object({
elements: z.array(z.discriminatedUnion('type', [
z.object({
type: z.literal('shape'),
kind: z.enum(['rectangle', 'circle', 'ellipse', 'triangle']),
position: zPoint,
width: z.number(),
height: z.number(),
fillColor: z.string(),
strokeColor: z.string().optional(),
strokeWidth: z.number().min(0).optional(),
borderRadius: z.number().optional(),
text: z.string(),
textAlignment: z.enum(['left', 'center', 'right']).optional(),
textColor: z.string().optional(),
textBold: z.boolean().optional(),
textItalic: z.boolean().optional(),
}),
z.object({
type: z.literal('path'),
kind: z.enum(['line', 'polyline']),
position: zPoint,
points: z.array(zPoint),
strokeColor: z.string(),
textAlignment: z.enum(['arrow-head-line']).optional(),
}),
z.object({
type: z.literal('image'),
mxc: z.string(),
filename: z.string(),
mimeType: z.enum(['image/gif', 'image/jpeg', 'image/png', 'image/svg+xml']),
position: zPoint,
width: z.number().min(0),
height: z.number().min(0),
}),
])),
})),
}),
});
And this is how to use it in NodeJS, Bun and Deno:
import fs from 'node:fs';
// Beware, these may throw errors
const whiteboard = JSON.parse(fs.readFileSync('my-template.nwb', 'utf8'));
const safeWhiteboard = zWhiteboard.parse(data);
console.log(safeWhiteboard);