Experimental!!!
Attempts to infer a typescript type by inspecting runtime javascript objects.
- You have a large API that is not currently typed
- There are no/poor specifications or documentation of what the shape of the objects it returns
- The responses it return vary, and it's difficult to know that you have a complete understanding of what it can return
- You want to write a type for it
The idea is that you want to run this on a real-world system, observing the real-world data for a while, then after some time has passed and you've observed many responses, you ask it to suggest a type that will match for all of them.
For each object you'd want to generate a type for, you'd create an instance of TypeGenerator
then call observe
, passing the data that you're wanting to generate a type for. After some time has passed and you think you have a complete understanding of what the API does, you'd call suggest
and it'll make you a typescript
type that fits the observed data.
npm install @itaylor/just-my-type
import TypeGenerator from '@itaylor/just_my_type';
import TypeGenerator from 'https://deno.land/x/just_my_type@0.0.1/main.ts';
const tg = new TypeGenerator('MyAPI');
// Assumption: this api is being called by your application somewhere on a regular basis.
export async function myApi() {
const data = await (await fetch('https://someserver.local/myApiThatINeedATypeFor/')).json();
tg.observe(data);
return data;
}
// This is a bad way to do this, but illustrates the point that `observe` should be called multiple times
// With all the different variations that need to be observed before calling `suggest`
setInterval(() => console.log(tg.suggest()), 5 * 60 * 1000);
Every 5 minutes you'd see something like this in the logs that would be the current representation of all the data that has been observed:
export type MyAPI = {
startIndex: number,
maxReturn: number,
items: Array<{
type: string,
children: Array<{
x: number,
y: number
}>
} | {
type: string,
messageText: string
} | {
type: string,
responses: Array<string>
}>
}