sia.hackernoon.com

I have always felt a bit annoyed whenever I’m working on a project and I have to add types to an api response.

The process usually goes like this:

write a function to fetch data
call the function on a component or page
either log the response or inspect the network tab to view the response
copy this response, create a new file and paste the response
start editing the response to create a new type or interface
import the new type to the function file and set it’s return type.

This had been going on for a while and then I thought to find a way to automate it.

This led to my discovery of a few libraries that can generate types. Unfortunately they were not convenient to work with because they were not generating the types effortlessly and in realtime. They required a lot of manual process as well. I wanted something I could just plug in and be done with and continue with my normal development flow.

So I decided to build a library that will automate this process for me.

realtime-api-types

The aim of this library is to take the pain away from trying to add types from your api responses.

With this library, as you get responses from an api, it generates the type for you and saves it to your project, in real time.

It also imports the new type and sets it as the return type of your api call.

Just import the helper function and start the server. Thats all. You can then go on with your normal development flow.

This setup is meant only for development and not for production! You should not deploy this into your production pipeline.

This library was inspired by an old project I stumbled upon, called MakeTypes.

Installation

To get started, simply run:

npm install realtime-api-types --save-dev

Configuration

There are some configs you need to set. Go to your package.json and add the following:

"realtime-api-types": {
"objectType": "type",
"typePath": "src/types",
"apiPath": "src/apis",
"fetchType": "axios"
}

The shape of the config is:

export type Config = {
 objectType: 'interface' | 'type'; // whether you want to generate interfaces or types
 fetchType: 'fetch' | 'axios'; // how you fetch data, fetch or axios
 typePath: string; // the path to where you want to save generated types
 apiPath: string; // the path to where your api methods exist
}

Start Service

To start the type generator service, run:

npx realtime-api-types - init

Code Sample & Usage

Wrap your apis in an object and wrap the object with typedApiWrapper.

You should make your api methods pure, simply return the api call.

// Example
// src/apis/exercise.ts api file
import {typedApiWrapper} from "realtime-api-types"
import axios from 'axios'

export const ExerciseApi = typedApiWrapper({
  // with fetch, property assignment style
  getExercises: () => fetch("https://example-api.com").then((res) => res.json()),
  // with axios, method style
  getExerciseById(id: string) {
    return axios.get(`https://example-api.com/${id}`);
  },
  // api post method
  postExercise(data: any) {
  return axios.post(`https://example-api.com`, data)
  }
});

// src/App.tsx file
useEffect(() => {
  ExerciseApi.getExercises()
}, [])

With this setup, whenever an api is called at any point in time, the service will intercept and try to generate type from the response.

When type generation is successful, the example file above would be automatically updated to something like this:

// updated exercise.ts api file
import {typedApiWrapper} from "realtime-api-types"
import axios from 'axios'
import { GetExercises } from "../types/getExercises";
import { GetExerciseById } from "../types/getExerciseById";
import { PostExercise } from "../types/postExercise";

export const ExerciseApi = typedApiWrapper({
  // with fetch, property assignment style
  getExercises: (): Promise<GetExercises> => fetch("https://example-api.com").then((res) => res.json()),
  // with axios, method style
  getExerciseById(id: string): Promise<{ data: GetExerciseById }> {
    return axios.get(`https://example-api.com/${id}`);
  },
  // api post method
  postExercise(data: any): Promise<{ data: PostExercise }> {
    return axios.post(`https://example-api.com`, data)
  }
});

Naming Convention

Note the naming convention in the example above.

The name of the type file is the same as the name of the api method called.
The name of the type itself is the same as the name of the api method but in pascal case.

React Native or Expo

For this to work with React native or Expo, make sure you follow their guide on how to enable api calls to localhost

Limitations

Cannot generate enums from response.
Cannot extend type from different type files. If the response from a call contains object that is similar to another type in another file, it cannot extend it. A new type will be generated in the new file.
Cannot give custom file names or type names. File name and type name is solely based on the name of the api method.
Once a type has been generated for an api call, the type will not update with new api calls. You need to delete the previous generated type file to generate a new one.

This library only serves to help you get started quickly and reduce time spent adding types to api calls. You might need to make some updates to the generated types sometimes. It does not solve all your type problems.

Let me know what you think and please feel free to contribute to the repo at https://github.com/ifeoluwak/realtime-api-types

npm — https://www.npmjs.com/package/realtime-api-types

github — https://github.com/ifeoluwak/realtime-api-types

Take the Pain Away From Adding Types From Your API Responses With This Automated Library

realtime-api-types