Flux is an application design paradigm used as a replacement for the more traditional MVC pattern. It is not a framework or a library but a new kind of architecture that complements React and the concept of Unidirectional Data Flow. Facebook uses this pattern internally when working with React.
The workflow between dispatcher, stores and views components with distinct inputs and outputs as follows:
Redux is a predictable state container for JavaScript apps based on the Flux design pattern. Redux can be used together with React, or with any other view library. It is tiny (about 2kB) and has no dependencies.
Redux follows three fundamental principles:
Instead of saying downsides we can say that there are few compromises of using Redux over Flux. Those are as follows:
redux-immutable-state-invariant
, Immutable.js, or instructing your team to write non-mutating code.mapStateToProps()
is a utility which helps your component get updated state (which is updated by some other components):
const mapStateToProps = (state) => {
return {
todos: getVisibleTodos(state.todos, state.visibilityFilter)
}
}
mapDispatchToProps()
is a utility which will help your component to fire an action event (dispatching action which may cause change of application state):
const mapDispatchToProps = (dispatch) => {
return {
onTodoClick: (id) => {
dispatch(toggleTodo(id))
}
}
}
It is recommended to always use the “object shorthand” form for the mapDispatchToProps
.
Redux wraps it in another function that looks like (…args) => dispatch(onTodoClick(…args)), and pass that wrapper function as a prop to your component.
const mapDispatchToProps = ({
onTodoClick
})
Dispatching an action within a reducer is an anti-pattern. Your reducer should be without side effects, simply digesting the action payload and returning a new state object. Adding listeners and dispatching actions within the reducer can lead to chained actions and other side effects.
You just need to export the store from the module where it created with createStore()
. Also, it shouldn't pollute the global window object.
store = createStore(myReducer)
export default store
DOM manipulation is very expensive which causes applications to behave slow and inefficient.
Due to circular dependencies, a complicated model was created around models and views.
Lot of data changes happens for collaborative applications(like Google Docs).
No way to do undo (travel back in time) easily without adding so much extra code.
These libraries are very different for very different purposes, but there are some vague similarities.
Redux is a tool for managing state throughout the application. It is usually used as an architecture for UIs. Think of it as an alternative to (half of) Angular. RxJS is a reactive programming library. It is usually used as a tool to accomplish asynchronous tasks in JavaScript. Think of it as an alternative to Promises. Redux uses the Reactive paradigm because the Store is reactive. The Store observes actions from a distance, and changes itself. RxJS also uses the Reactive paradigm, but instead of being an architecture, it gives you basic building blocks, Observables, to accomplish this pattern.
You can dispatch an action in componentDidMount()
method and in render()
method you can verify the data.
class App extends Component {
componentDidMount() {
this.props.fetchData()
}
render() {
return this.props.isLoaded
? <div>{'Loaded'}</div>
: <div>{'Not Loaded'}</div>
}
}
const mapStateToProps = (state) => ({
isLoaded: state.isLoaded
})
const mapDispatchToProps = { fetchData }
export default connect(mapStateToProps, mapDispatchToProps)(App)
You need to follow two steps to use your store in your container:
mapStateToProps()
: It maps the state variables from your store to the props that you specify.mapStateToProps
function is connected to the container. You can import connect()
from react-redux
.import React from 'react'
import { connect } from 'react-redux'
class App extends React.Component {
render() {
return <div>{this.props.containerData}</div>
}
}
function mapStateToProps(state) {
return { containerData: state.data }
}
export default connect(mapStateToProps)(App)
You need to write a root reducer in your application which delegate handling the action to the reducer generated by combineReducers()
.
For example, let us take rootReducer()
to return the initial state after USER_LOGOUT
action. As we know, reducers are supposed to return the initial state when they are called with undefined
as the first argument, no matter the action.
const appReducer = combineReducers({
/* your app's top-level reducers */
})
const rootReducer = (state, action) => {
if (action.type === 'USER_LOGOUT') {
state = undefined
}
return appReducer(state, action)
}
In case of using redux-persist
, you may also need to clean your storage. redux-persist
keeps a copy of your state in a storage engine. First, you need to import the appropriate storage engine and then, to parse the state before setting it to undefined and clean each storage state key.
const appReducer = combineReducers({
/* your app's top-level reducers */
})
const rootReducer = (state, action) => {
if (action.type === 'USER_LOGOUT') {
Object.keys(state).forEach(key => {
storage.removeItem(`persist:${key}`)
})
state = undefined
}
return appReducer(state, action)
}
The @ symbol is in fact a JavaScript expression used to signify decorators. Decorators make it possible to annotate and modify classes and properties at design time.
Let's take an example setting up Redux without and with a decorator.
import React from 'react'
import * as actionCreators from './actionCreators'
import { bindActionCreators } from 'redux'
import { connect } from 'react-redux'
function mapStateToProps(state) {
return { todos: state.todos }
}
function mapDispatchToProps(dispatch) {
return { actions: bindActionCreators(actionCreators, dispatch) }
}
class MyApp extends React.Component {
// ...define your main app here
}
export default connect(mapStateToProps, mapDispatchToProps)(MyApp)
import React from 'react'
import * as actionCreators from './actionCreators'
import { bindActionCreators } from 'redux'
import { connect } from 'react-redux'
function mapStateToProps(state) {
return { todos: state.todos }
}
function mapDispatchToProps(dispatch) {
return { actions: bindActionCreators(actionCreators, dispatch) }
}
@connect(mapStateToProps, mapDispatchToProps)
export default class MyApp extends React.Component {
// ...define your main app here
}
The above examples are almost similar except the usage of decorator. The decorator syntax isn't built into any JavaScript runtimes yet, and is still experimental and subject to change. You can use babel for the decorators support.
You can use Context in your application directly and is going to be great for passing down data to deeply nested components which what it was designed for.
Whereas Redux is much more powerful and provides a large number of features that the Context API doesn't provide. Also, React Redux uses context internally but it doesn't expose this fact in the public API.
Reducers always return the accumulation of the state (based on all previous and current actions). Therefore, they act as a reducer of state. Each time a Redux reducer is called, the state and action are passed as parameters. This state is then reduced (or accumulated) based on the action, and then the next state is returned. You could reduce a collection of actions and an initial state (of the store) on which to perform these actions to get the resulting final state.
You can use redux-thunk
middleware which allows you to define async actions.
Let's take an example of fetching specific account as an AJAX call using fetch API:
export function fetchAccount(id) {
return dispatch => {
dispatch(setLoadingAccountState()) // Show a loading spinner
fetch(`/account/${id}`, (response) => {
dispatch(doneFetchingAccount()) // Hide loading spinner
if (response.status === 200) {
dispatch(setAccount(response.json)) // Use a normal function to set the received state
} else {
dispatch(someError)
}
})
}
}
function setAccount(data) {
return { type: 'SET_Account', data: data }
}
Keep your data in the Redux store, and the UI related state internally in the component.
The best way to access your store in a component is to use the connect()
function, that creates a new component that wraps around your existing one. This pattern is called Higher-Order Components, and is generally the preferred way of extending a component's functionality in React. This allows you to map state and action creators to your component, and have them passed in automatically as your store updates.
Let's take an example of <FilterLink>
component using connect:
import { connect } from 'react-redux'
import { setVisibilityFilter } from '../actions'
import Link from '../components/Link'
const mapStateToProps = (state, ownProps) => ({
active: ownProps.filter === state.visibilityFilter
})
const mapDispatchToProps = (dispatch, ownProps) => ({
onClick: () => dispatch(setVisibilityFilter(ownProps.filter))
})
const FilterLink = connect(
mapStateToProps,
mapDispatchToProps
)(Link)
export default FilterLink
Due to it having quite a few performance optimizations and generally being less likely to cause bugs, the Redux developers almost always recommend using connect()
over accessing the store directly (using context API).
class MyComponent {
someMethod() {
doSomethingWith(this.context.store)
}
}
Component is a class or function component that describes the presentational part of your application.
Container is an informal term for a component that is connected to a Redux store. Containers subscribe to Redux state updates and dispatch actions, and they usually don't render DOM elements; they delegate rendering to presentational child components.
Constants allows you to easily find all usages of that specific functionality across the project when you use an IDE. It also prevents you from introducing silly bugs caused by typos – in which case, you will get a ReferenceError
immediately.
Normally we will save them in a single file (constants.js
or actionTypes.js
).
export const ADD_TODO = 'ADD_TODO'
export const DELETE_TODO = 'DELETE_TODO'
export const EDIT_TODO = 'EDIT_TODO'
export const COMPLETE_TODO = 'COMPLETE_TODO'
export const COMPLETE_ALL = 'COMPLETE_ALL'
export const CLEAR_COMPLETED = 'CLEAR_COMPLETED'
In Redux, you use them in two places:
Let's take actions.js
:
import { ADD_TODO } from './actionTypes';
export function addTodo(text) {
return { type: ADD_TODO, text }
}
Let's create reducer.js
:
import { ADD_TODO } from './actionTypes'
export default (state = [], action) => {
switch (action.type) {
case ADD_TODO:
return [
...state,
{
text: action.text,
completed: false
}
];
default:
return state
}
}
There are a few ways of binding action creators to dispatch()
in mapDispatchToProps()
.
Below are the possible options:
const mapDispatchToProps = (dispatch) => ({
action: () => dispatch(action())
})
const mapDispatchToProps = (dispatch) => ({
action: bindActionCreators(action, dispatch)
})
const mapDispatchToProps = { action }
The third option is just a shorthand for the first one.
If the ownProps
parameter is specified, React Redux will pass the props that were passed to the component into your connect functions. So, if you use a connected component:
import ConnectedComponent from './containers/ConnectedComponent';
<ConnectedComponent user={'john'} />
The ownProps
inside your mapStateToProps()
and mapDispatchToProps()
functions will be an object:
{ user: 'john' }
You can use this object to decide what to return from those functions.
Most of the applications has several top-level directories as below:
This structure works well for small and medium size apps.
redux-saga
is a library that aims to make side effects (asynchronous things like data fetching and impure things like accessing the browser cache) in React/Redux applications easier and better.
It is available in NPM:
$ npm install --save redux-saga
Saga is like a separate thread in your application, that's solely responsible for side effects. redux-saga
is a redux middleware, which means this thread can be started, paused and cancelled from the main application with normal Redux actions, it has access to the full Redux application state and it can dispatch Redux actions as well.
Both call()
and put()
are effect creator functions. call()
function is used to create effect description, which instructs middleware to call the promise. put()
function creates an effect, which instructs middleware to dispatch an action to the store.
Let's take example of how these effects work for fetching particular user data.
function* fetchUserSaga(action) {
// `call` function accepts rest arguments, which will be passed to `api.fetchUser` function.
// Instructing middleware to call promise, it resolved value will be assigned to `userData` variable
const userData = yield call(api.fetchUser, action.userId)
// Instructing middleware to dispatch corresponding action.
yield put({
type: 'FETCH_USER_SUCCESS',
userData
})
}
Redux Thunk middleware allows you to write action creators that return a function instead of an action. The thunk can be used to delay the dispatch of an action, or to dispatch only if a certain condition is met. The inner function receives the store methods dispatch()
and getState()
as parameters.
Both Redux Thunk and Redux Saga take care of dealing with side effects. In most of the scenarios, Thunk uses Promises to deal with them, whereas Saga uses Generators. Thunk is simple to use and Promises are familiar to many developers, Sagas/Generators are more powerful but you will need to learn them. But both middleware can coexist, so you can start with Thunks and introduce Sagas when/if you need them.
Redux DevTools is a live-editing time travel environment for Redux with hot reloading, action replay, and customizable UI. If you don't want to bother with installing Redux DevTools and integrating it into your project, consider using Redux DevTools Extension for Chrome and Firefox.
Some of the main features of Redux DevTools are below:
persistState()
store enhancer, you can persist debug sessions across page reloads.Selectors are functions that take Redux state as an argument and return some data to pass to the component.
For example, to get user details from the state:
const getUserData = state => state.user.data
These selectors have two main benefits,
Redux Form works with React and Redux to enable a form in React to use Redux to store all of its state. Redux Form can be used with raw HTML5 inputs, but it also works very well with common UI frameworks like Material UI, React Widgets and React Bootstrap.
Some of the main features of Redux Form are:
You can use applyMiddleware()
.
For example, you can add redux-thunk
and logger
passing them as arguments to applyMiddleware()
:
import { createStore, applyMiddleware } from 'redux'
const createStoreWithMiddleware = applyMiddleware(ReduxThunk, logger)(createStore)
You need to pass initial state as second argument to createStore:
const rootReducer = combineReducers({
todos: todos,
visibilityFilter: visibilityFilter
})
const initialState = {
todos: [{ id: 123, name: 'example', completed: false }]
}
const store = createStore(
rootReducer,
initialState
)
Relay is similar to Redux in that they both use a single store. The main difference is that relay only manages state originated from the server, and all access to the state is used via GraphQL queries (for reading data) and mutations (for changing data). Relay caches the data for you and optimizes data fetching for you, by fetching only changed data and nothing more.
Actions are plain JavaScript objects or payloads of information that send data from your application to your store. They are the only source of information for the store. Actions must have a type property that indicates the type of action being performed.
For example, let's take an action which represents adding a new todo item:
{
type: ADD_TODO,
text: 'Add todo item'
}