GithubHelp home page GithubHelp logo

react-hooks-redux-thunk-readme's Introduction

Asynchronous Web Requests with Using Thunk and Redux

Learning Goals

  • Use action creator functions to make asynchronous web requests
  • Explain why middleware is necessary for asynchronous actions
  • Use the redux-thunk middleware for asynchronous actions

Introduction

Part of the value of using Redux is that it provides a centralized way to control the data of an application. In a standard React/Redux application, any child component can connect to the store directly from anywhere in the app. This allows us to keep many of our React components simple — no need for passing props through many nested components, no need to use component state to keep track of all the data. A lot of code that would normally be stored in React components can be removed or replaced.

With Redux, we can focus more on presentation in our React components, and use actions and reducers to handle the logic of organizing data. In following with this pattern, we'll be discussing a package that works in conjunction with Redux: redux-thunk.

redux-thunk handles asynchronous calls when working with Redux. Think for a moment: we have Redux handling all our app's data. So far, it's all been hard-coded data, i.e. data that we set ourselves. It would be great if we could start getting data from other sources.

Well, if we had a server or an API, we could fetch some remote data, but we're presented with a familiar problem: we've just removed a lot of logic from our components and now we're going to add more logic? Specifically, we're going to fetch data we'll likely want to keep in our Redux store — adding code to our components seems to be a step backwards.

With redux-thunk, we can incorporate asynchronous code in with our Redux actions. This allows us to continue keeping our components relatively simple and more focused on presentation. In this lesson, we're going to go through what Thunk is and how it is implemented with Redux.

Trying to Send an Asynchronous Request in Redux

We're familiar with the Redux pattern in which the store dispatches an action to the reducer, the reducer uses that action to make changes to the state, and components re-render with new data.

Going back to hard-coded examples, in previous lessons, we populated our store using data inside an action creator function. Something like this:

function fetchAstronauts() {
  const astronauts = [
    { name: "Neil Armstrong", craft: "Apollo 11" },
    { name: "Buzz Aldrin", craft: "Apollo 11" },
    { name: "Michael Collins", craft: "Apollo 11" },
  ];
  return {
    type: "astronauts/add",
    astronauts,
  };
}

What happens though, when we're ready to pull in real live data from an external source like an API?

Well, we already know how to make a web request. We can use something like JavaScript's native Fetch API to send a web request:

fetch("http://api.open-notify.org/astros.json");

So, can we simply make a fetch request inside our action creator function instead of hard-coding our data? The code below is a good attempt, but it ultimately ends in failure and disappointment:

// ./src/features/astronauts/Astronauts.js
import React from "react";
import { useDispatch, useSelector } from "react-redux";
import { fetchAstronauts } from "./astronautsSlice";

function Astronauts() {
  const dispatch = useDispatch();

  const astronauts = useSelector((state) => state.astronauts.entities);

  function handleClick() {
    // dispatch the action creator (see below!)
    dispatch(fetchAstronauts());
  }

  const astronautsList = astronauts.map((astro) => (
    <li key={astro.id}>{astro.name}</li>
  ));

  return (
    <div>
      <button onClick={handleClick}>Get Astronauts</button>
      {astronautsList}
    </div>
  );
}

export default Astronauts;
// ./src/features/astronauts/astronautsSlice.js
// Action Creators
export function fetchAstronauts(astronauts) {
  return {
    type: "astronauts/astronautsLoaded",
    payload: astronauts,
  };
}

// Reducers
const initialState = {
  entities: [], //array of astronauts
};

export default function reducer(state = initialState, action) {
  switch (action.type) {
    case "astronauts/astronautsLoaded":
      return {
        ...state,
        entities: action.payload,
      };

    default:
      return state;
  }
}

So if you look at the code above, you get a sense for what we are trying to do. When a user clicks on the button, we call the handleOnClick() function. This calls our action creator, the fetchAstronauts() function. The action creator then hits the API, and returns an action with our data, which then updates the state through the reducer.

While this might seem like it should work, in reality we have a big problem.

Fetch requests in JavaScript are asynchronous. That means if we make a fetch request at the first line of our fetchAstronauts() function:

export function fetchAstronauts() {
  const astronauts = fetch("http://api.open-notify.org/astros.json");
  return {
    type: "astronauts/astronautsLoaded",
    payload: astronauts,
  };
}

The code on the second line will start running before the web request resolves and we have a response that we can work with.

A fetch() request returns something called a Promise. A Promise object is an object that represents some value that will be available later. We can access the data when the promise "resolves" and becomes available by chaining a then() function onto our fetch() call.

export function fetchAstronauts() {
  const astronauts = fetch("http://api.open-notify.org/astros.json").then(
    (response) => response.json()
  );
  return {
    type: "astronauts/astronautsLoaded",
    payload: astronauts,
  };
}

Our then() function will run when the Promise that fetch() returns is resolved, allowing us to access the response data and parse it into JSON. This doesn't solve our problem though because the fetchAstronauts() function will still return before the Promise is resolved.

There's another problem. Because retrieving data takes time, and because we always want our Redux store to reflect the current application state, we want to represent the state of the application in between the user asking for data and the application receiving the data. It's almost like each time a user asks for data we want to dispatch two actions to update our state: one to place our state as loading, and another to update the state with the data.

So these are the steps we want to happen when the user wishes to call the API:

  1. Invoke fetchAstronauts()
  2. Directly after invoking fetchAstronauts() dispatch an action to indicate that we are loading data.
  3. Call the fetch() method, which runs, and returns a Promise that we are waiting to resolve.
  4. When the Promise resolves, dispatch another action with a payload of the fetched data that gets sent to the reducer.

Great. So how do we do all of this?

We Need Middleware

So we need a way to dispatch an action saying we are loading data, then to make a request to the API, and then to wait for the response and then dispatch another action with the response data.

Lucky for us, we can use some middleware for exactly that! Middleware, in this case, will allow us to slightly alter the behavior of our actions, allowing us to add in asynchronous requests. In this case, for middleware, we'll be using Thunk.

To use redux-thunk you would need to install the NPM package:

$ npm install redux-thunk

Then, when you initialize the store in your index.js file, you can incorporate your middleware like this:

// src/index.js

import React from "react";
import ReactDOM from "react-dom";
import { Provider } from "react-redux";
import { createStore, applyMiddleware } from "redux";
import thunkMiddleware from "redux-thunk";
import App from "./App";
import rootReducer from "./reducer";

const store = createStore(rootReducer, applyMiddleware(thunkMiddleware));

ReactDOM.render(
  <Provider store={store}>
    <App />
  </Provider>,
  document.getElementById("root")
);

Notice that we imported in a new function applyMiddleware() from redux, along with thunk from the redux-thunk package, and passed in applyMiddleware(thunk) as a second argument to createStore.

We can also set up our app to use the Redux DevTools. We'll use another package, redux-devtools-extension, to help with the setup:

$ npm install redux-devtools-extension

Then update our index.js file like so:

import React from "react";
import ReactDOM from "react-dom";
import { Provider } from "react-redux";
import { createStore, applyMiddleware } from "redux";
import thunkMiddleware from "redux-thunk";
import { composeWithDevTools } from "redux-devtools-extension";
import App from "./App";
import rootReducer from "./reducer";

const composedEnhancer = composeWithDevTools(applyMiddleware(thunkMiddleware));

const store = createStore(rootReducer, composedEnhancer);

ReactDOM.render(
  <Provider store={store}>
    <App />
  </Provider>,
  document.getElementById("root")
);

Using Redux-Thunk Middleware

In the above code, we tell our store to use the redux-thunk middleware. This middleware will do a couple of interesting things:

  1. redux-thunk allows us to return a function inside of our action creator. Normally, our action creator returns a plain JavaScript object, so returning a function is a pretty big change.
  2. That function receives the store's dispatch function as its argument. With that, we can dispatch multiple actions from inside that returned function.

Let's see the code and then we'll walk through it.

// src/features/astronauts/astronautsSlice.js
export function fetchAstronauts() {
  return function (dispatch) {
    // dispatch an initial action to set up a "loading" state
    dispatch({ type: "astronauts/astronautsLoading" });

    // initiate a network request with fetch
    fetch("http://api.open-notify.org/astros.json")
      .then((response) => response.json())
      .then((astronauts) =>
        // when we have data from the response, dispatch another action to add the data to our Redux store
        dispatch({
          type: "astronauts/astronautsLoaded",
          payload: astronauts.people,
        })
      );
  };
}

You can see above that we are returning a function and not an action, and that the power we now get is the ability to dispatch actions from inside of the returned function.

With that power, we first dispatch an action to indicate that we are about to make a request to our API. Then we make the request. We do not hit our then() function until the response is received, which means that we are not dispatching our action of type 'astronauts/astronautsLoaded' until we receive our data. Thus, we are able to send along that data.

Reviewing Everything Together

Let's review the whole application now with redux-thunk configured. First we have index.js, which now imports thunkMiddleware and applyMiddleware and uses them when creating the Redux store:

// ./src/index.js
import React from "react";
import ReactDOM from "react-dom";
import { Provider } from "react-redux";
import { createStore, applyMiddleware } from "redux";
import thunkMiddleware from "redux-thunk";
import { composeWithDevTools } from "redux-devtools-extension";
import App from "./App";
import rootReducer from "./reducers";

const composedEnhancer = composeWithDevTools(applyMiddleware(thunkMiddleware));

const store = createStore(rootReducer, composedEnhancer);

ReactDOM.render(
  <Provider store={store}>
    <App />
  </Provider>,
  document.getElementById("root")
);

The Astronauts component we showed earlier can remain the same — note that although we've called a function fetchAstronauts(), no actual asynchronous code is in the component. The component's main purpose is to render JSX. It uses data from Redux via useSelector() and connects an onClick event to an action through useDispatch():

// ./src/features/astronauts/Astronauts.js
import React from "react";
import { useDispatch, useSelector } from "react-redux";
import { fetchAstronauts } from "./astronautsSlice";

function Astronauts() {
  const dispatch = useDispatch();

  const astronauts = useSelector((state) => state.astronauts.entities);

  function handleClick() {
    dispatch(fetchAstronauts());
  }

  const astronautsList = astronauts.map((astro) => (
    <li key={astro.id}>{astro.name}</li>
  ));

  return (
    <div>
      <button onClick={handleClick}>Get Astronauts</button>
      {astronautsList}
    </div>
  );
}

export default Astronauts;

What happens when the onClick event is fired? All of that logic is taken care of outside of the component, in our fetchAstronauts() action:

// src/features/astronauts/astronautsSlice.js
export function fetchAstronauts() {
  return function (dispatch) {
    dispatch({ type: "astronauts/astronautsLoading" });
    fetch("http://api.open-notify.org/astros.json")
      .then((response) => response.json())
      .then((astronauts) =>
        dispatch({
          type: "astronauts/astronautsLoaded",
          payload: astronauts.people,
        })
      );
  };
}

With redux-thunk configured, our actions can now return a function. We must write the function, but we know that dispatch() is passed in as an argument. Notice in the code above that there are two calls to dispatch(), first passing in { type: 'astronauts/astronautsLoading' } before the fetch() call, then passing in { type: 'astronauts/astronautsLoaded', payload: astronauts } inside .then(). By having both dispatch() calls, it is possible to know just before our application sends a remote request, and then immediately after that request is resolved.

We can update our reducer to include both types and to also change a bit of state to indicate if data is in the process of being fetched. We'll modify the initial state to do this:

// ./src/features/astronauts/astronautsSlice.js
const initialState = {
  entities: [], //array of astronauts
  status: "idle", //loading status for fetch
};

export default function reducer(state = initialState, action) {
  switch (action.type) {
    case "astronauts/astronautsLoaded":
      return {
        ...state,
        status: "idle",
        entities: action.payload,
      };
    case "astronauts/astronautsLoading":
      return {
        ...state,
        status: "loading",
      };
    default:
      return state;
  }
}

Now, we have a way to indicate in our app when data is being loaded! If status is 'loading', we could display a loading message in our components.

Conclusion

We saw that when retrieving data from APIs, we run into a problem where the action creator returns an action before the data is retrieved. To resolve this, we use a middleware called redux-thunk. redux-thunk allows us to return a function inside of our action creator instead of a plain JavaScript object. That returned function receives the store's dispatch function, and with that we are able to dispatch multiple actions: one to place the state in a loading state, and another to update our store with the returned data.

react-hooks-redux-thunk-readme's People

Contributors

ihollander avatar lizbur10 avatar lukeghenco avatar jeffkatzy avatar sophiedebenedetto avatar maxwellbenton avatar gj avatar achasveachas avatar scrabill avatar sdcrouse avatar gizemcdeniz avatar allisonkadel avatar walwoodr avatar mdo5004 avatar jmburges avatar jjseymour avatar crwhitesides avatar beejluig avatar

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google ❤️ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.