By the end of this workshop, you will be familiar with the basic techniques used to build user interfaces in React.
This workshop will take you step-by-step through the process of implementing the design and functional requirements for a (wait for it) TO-DO APP!!!! Good news: all the CSS has already been written for you, so you can focus on writing just JavaScript!
Here's how your killer To-do app will look:
This workshop is inspired by the "Thinking in React" tutorial.
$ git clone git@github.com:cjcenizal/react-design-workshop.git
$ cd react-design-workshop
$ npm install
$ npm start
Each of the challenges below has a screenshot of what you'll be building, instructions on how to tackle the challenge (including code snippets and links to external resources), and a summary of key take-aways. When you're done with these main challenges, consider trying out some of the bonus challenges that follow.
Your first challenge is to render a list of To-dos inside of a panel.
Open up src/ToDoApp.js
. This is your To-do app, where you'll do most of your work. It's just a
React component, typified by the render
method. If you look in the browser, you'll see that React
uses the render method to determine what to draw to the screen. Feel free to make some changes to
what's returned by the render method before moving on.
Let's start by changing this render method to render a Panel. This is how we do that:
render() {
return (
<Panel>
</Panel>
);
}
Panel is a React component that will be responsible for drawing this part of our UI. But Panel isn't
available in our current scope yet. So let's import it! It makes sense to put all of our reusable components in
one place, so let's plan on putting them all in a /components
directory.
import {
Panel,
} from './components';
Now that we've imported it, we need to create it and export it. Let's create it first. Create a file
called /components/Panel.js
. Inside of it, export the Panel component:
import React, {
PropTypes,
} from 'react';
export const Panel = props => {
return (
<div className="panel">
{props.children}
</div>
);
};
Panel.propTypes = {
children: PropTypes.node,
};
We're doing a few important things here:
- We're importing React. This is required anywhere you write JSX.
- We're exporting a stateless functional component. This is the simplest kind of component you can create in React.
- We're composing the children property into our component. This is the simplest way to compose React components.
- We're also explicitly stating that
children
can only be of typenode
, which is defined by React. If an engineer tries to pass a non-node as a child (e.g. a function), you'll see an error in the console.
We need to do one more thing! Let's create an index file for exporting all of the beautiful components
we're going to build. Create /components/index.js
and export the Panel from it:
export { Panel } from './Panel';
Now, in the browser, you should see your Panel rendered to the screen!
We mentioned how React will verify the props you pass to a component based on the PropTypes you've defined. Let's verify this by passing a function to Panel as its children:
render() {
return (
<Panel>
{() => {}}
</Panel>
);
}
In the browser you should see an error like this:
Warning: Failed prop type: Invalid prop `children` supplied to `Panel`, expected a ReactNode.
Success! Well, failure. But good failure! You know what went wrong and you know how to fix it. So let's do that.
We talked about children in the previous steps. Let's make this Panel a daddy! (Or mommy, depending on how your Panel gender-identifies.) In the render method, let's swap out that unacceptable function with some acceptable nodes:
render() {
return (
<Panel>
<h1 className="panelHeader">
To-dos
</h1>
</Panel>
);
}
Looks like HTML doesn't it? The only difference is we have to use the word "className" instead of "class" to style our elements.
This renders what we want it to, but we also want to keep creating reusable components. Let's extract this "markup" into a new component. Follow the same process we established for Panel, but create a PanelHeader component:
import React, {
PropTypes,
} from 'react';
export const PanelHeader = props => {
return (
<h1 className="panelHeader">
{props.children}
</h1>
);
};
PanelHeader.propTypes = {
children: PropTypes.node,
};
And then swap it in for our markup inside our original render method:
render() {
return (
<Panel>
<PanelHeader>
To-dos
</PanelHeader>
</Panel>
);
}
See how the string "To-dos" gets rendered in the browser?
It's because we're providing that string as the children
prop to the PanelHeader component. Try changing
it to a different value and see it change in the browser.
Take-aways
- Components should generally be designed to be reusable.
- Components lend themselves to composition.
- A good process for creating components is to write "markup" first, and then extract components, second.
- Use stateless functional components to create simple components.
- Use PropTypes to do basic type-checking on dependencies.
- Use an index.js file to export cohesive modules from one place (and import them from that one place).
Next up, we're going to add some To-dos to our Panel.
Let's begin by creating /components/ToDoList.js
and adding it to our render method. Let's see if you can take the original markup and use it to create the new component, based on what you've learned so far:
<ol class="toDoList">
<!-- The To-dos will go in here. -->
</ol>
Add it to your render method, thusly:
render() {
return (
<Panel>
<PanelHeader>
To-dos
</PanelHeader>
<ToDoList>
</ToDoList>
</Panel>
);
}
This doesn't really do much in the browser so let's add some items to it. Create /components/ToDoListemItem.js
so that it represents this markup:
<li class="toDoListItem">
<!-- Render the `children` prop in here. -->
</li>
Add three ToDoListItems to the ToDoList. Voila! You've got a real bonafide list of things to get done.
<ToDoList>
<ToDoListItem>
Build React app
</ToDoListItem>
<ToDoListItem>
???
</ToDoListItem>
<ToDoListItem>
Profit!
</ToDoListItem>
</ToDoList>
Now we're going to render these items dynamically by introducing the concept of state
. The state
object is a special object that allows you to change how
a component is rendered over time. This is because every time this object is changed, React will
execute the render
method of that component, which will in turn execute the render
method of
its children, and so on. The result is a complete re-draw of that part of the DOM tree (there are
actually some optimizations
that occur under the hood, but that's another tangent).
It doesn't make a lot of sense to hard-code our ToDoListItems in the render method like this. We want to dynamically define the To-dos that get rendered. To do that (ha! "to do" that) we need to define an array on our state object which will contain these To-dos. We'll refer to this array in the render method.
In the constructor, define a state object after the super
invocation:
constructor() {
super();
this.state = {
toDos: [
'Build React app',
'???',
'Profit!',
],
};
}
Now let's wire up our render method to use this array. First, we'll extract and abstract the responsibility for rendering each To-do into its own method. As your React views grow, delegating responsibility for rendering different parts of the UI into their own render methods can help you scale the code:
renderToDos() {
return this.state.toDos.map(toDo => {
return (
<ToDoListItem>
{toDo}
</ToDoListItem>
);
});
}
render() {
return (
<Panel>
<PanelHeader>
To-dos
</PanelHeader>
<ToDoList>
{this.renderToDos()}
</ToDoList>
</Panel>
);
}
At this point you should see the To-dos rendered correctly in the browser, but in the console you should see this error:
Warning: Each child in an array or iterator should have a unique "key" prop. Check the render method of `ToDoApp`.
This is React complaining that it won't be able to optimize the way it renders your array of To-dos because it won't be able to tell which is which in the event that you reorder, remove, update, or add any.
We can fix this by adding a key
prop
to each ToDoListItem. Unfortunately, this is a hack and an anti-pattern. It will make the warning
go away but it's not the right long-term solution. We'll go into that in the next step.
return this.state.toDos.map((toDo, index) => {
return (
<ToDoListItem key={index}>
{toDo}
</ToDoListItem>
);
});
Keys should be used to reliably identify each individual element in an array. We can do this by
assigning IDs to our To-dos. We could update our array to be an array of objects instead of strings, and
assign an id
property to each object. But this would still mean we'd have to manually assign an
ID to each one. This also won't scale well as we need to associate more
data with each To-do, e.g. creation date, author, and severity. Let's solve these problems
by creating a ./models/ToDo.js
file.
let count = 0;
export class ToDo {
constructor(body) {
this.id = count++;
this.body = body;
}
}
Don't forget to create a /models/index.js
file to export this class, and then import it into your React
app:
import {
ToDo,
} from './models';
Now we can update our state object to contain instances of this class:
toDos: [
new ToDo('Build React app'),
new ToDo('???'),
new ToDo('Profit!'),
],
Can you figure out how to update your renderToDos
method to refer to the body
property on these instances?
Take-aways
- In React, user interfaces are composed of React components which are composed of other React components, all the way down.
- Data that changes should be assigned to the state object.
- Typically, this data is then passed to various children via their props by the render method.
- The render method can be split up into multiple methods for scalability.
- Assign a
key
prop to elements within an array, to help React identify them.
Time for things to get interesting! Let's let people actually create new To-dos. After all, there's always more... to do.
Our UI calls for a horizontal rule and a label for our form. Feel free to skip to the next section if you don't care about getting the UI just right, and just want to learn some React. For everyone else, here's what we want to add to our render method:
<HorizontalRule />
<Label>
Create new To-do
</Label>
These components need to represent the following markup:
<hr class="horizontalRule" />
<label class="label">
<!-- Render props.children here -->
</label>
Make it so!
Let's move on and build the UI that will drive our to-do-creation functionality. We need to give users a field
where they can type in the body of a new To-do. Create /components/TextInput.js
.
This component is going to be what the React team calls a controlled component. With basic HTML, form elements store their state directly in the DOM. But we're using React, which destroys and rebuilds the DOM all the time, and expects all state to be stored in JavaScript.
Check out the value
and onChange
props we define below. We're going to provide the value for
this input in our app's render method, and when the user changes this value, we're going to
update and re-provide the value to the component. From the user's point of view, they'll be
changing the value of the input as they type.
export const TextInput = props => {
return (
<input
type="text"
className="textInput"
value={props.value}
onChange={props.onChange}
/>
);
};
TextInput.propTypes = {
value: PropTypes.string,
onChange: PropTypes.func,
};
Back in our React app, instantiate the TextInput in the render method:
<TextInput
value={this.state.newToDoBody}
onChange={this.handleNewToDoInputChange}
/>
Now we need to close the loop by adding the newToDoBody
property to the state object, and by
adding the onChange handler, which will update this state.
handleNewToDoInputChange(event) {
this.setState({
newToDoBody: event.target.value,
});
}
When we call setState
we're telling React to change the state object and to
invoke the render
method again. Note that you can invoke setState multiple times in the same
call stack and React will still only invoke render once.
There's one last thing we need to do, and that's to bind this method to the scope of our To-do app. Do this at the end of the constructor:
this.handleNewToDoInputChange = this.handleNewToDoInputChange.bind(this);
You'll need to do this for every event handler you create during this workshop.
In your browser, you should now be able to type into the TextInput and it should behave as you expect.
Now let's add a button so we can add a new To-do to the list!
Create a file for the Button component. It should accept children
, onClick
, and isDisabled
props. This component should render a button element with the "button" class. Assign the onClick
prop
to its onClick
prop, and assign the isDisabled
prop to its disabled
prop.
Import this component into your app and add it to the UI. For now, only add an onClick callback:
<Button onClick={this.handleCreateToDoClick}>
Create To-do
</Button>
And define the handleCreateToDoClick
handler. Don't forget to bind it to the correct scope!
handleCreateToDoClick() {
this.createToDo();
}
Now we need to update our toDos array with a new one when the user clicks the button. Can you figure
out how to do that? Remember, you don't want to reach into the state object and mutate it directly.
You need to invoke setState
and provide it with new values for the properties you want to change.
Spoiler alert! Here's how to do it:
createToDo() {
this.setState({
toDos: this.state.toDos.concat(new ToDo(this.state.newToDoBody)),
newToDoBody: '',
});
}
It's kind of weird to let the user create To-dos with nothing in them. We should disable the button
when the input is empty. Let's add the isDisabled
prop to our Button instance and use a helper
method to figure out if we're in a state suitable for creating a new To-do. This is called deriving state.
canCreateNewToDo() {
return this.state.newToDoBody.trim().length > 0;
}
<Button
onClick={this.handleCreateToDoClick}
isDisabled={!this.canCreateNewToDo()}
>
Create To-do
</Button>
Not all state needs to go on the state
object. A good rule of thumb is to derive as much state at
render-time as you can. (You'll find exceptions to this rule, of course.)
Time to polish the user experience a bit. We want to create a new To-do by typing and hitting enter.
We can support this behavior by wrapping the form components in a <form>
.
<form onSubmit={this.handleCreateTodDoFormSubmit}>
{/* The TextInput and Button go in here. */}
</form>
Typically, submitting a form requires the page to reload, but we want to block that behavior. This is
what our onSubmit
event handler will do:
handleCreateTodDoFormSubmit(event) {
event.preventDefault();
}
Last reminder, don't forget to bind this handler in the constructor.
Note that React events are actually SyntheticEvents
,
and not native browser events, though you can generally use them the same way. Read the docs for the
full implications of this detail.
Take-aways
- Use controlled components to build form elements.
- Rememebr to bind event handlers to the component's scope.
- Call
setState
to modify the state object and cause React to re-render your component tree.- Think about which state needs to go on the
state
object and which can be derived state. In general, try to derive as much state as possible.- The React event system simulates the native browser event system, but is a fundamentally different beast.
OK, technically you've made a React app by this point, so let's take that one off the list. Crap! There's no functionality for deleting To-dos yet. Guess we know what we gotta do next.
Let's get the ToDoListItemButton
component out of the way. We'll use this to render the actual "Delete" button for each To-do.
Put this in its render method, and make sure it accepts the onClick
prop:
<button
className="toDoListItem__button"
onClick={props.onClick}
>
{props.children}
</button>
Now for the fun stuff. Our design calls for this button to be on the right side of each ToDoListItem.
Fortunately, because ToDoListItem just wraps the children
prop, we can add this functionality
without editing ToDoListItem's implementation details at all. This is a huge backwards-compatibility win! If we don't change the way ToDoListItem works, then
we don't risk breaking any part of our UI which depends on it. This is important in bigger apps.
Here's how we're going to do this. We're going to create a brand new component which composes
ToDoListItem
. Let's call it DeletableToDoListItem
. Go ahead and create this component right now.
Here's how it will look in all its glory:
import { ToDoListItem } from './ToDoListItem';
import { ToDoListItemButton } from './ToDoListItemButton';
export const DeletableToDoListItem = props => {
function onDeleteClick() {
props.onDeleteClick(props.id);
}
return (
<ToDoListItem>
{props.children}
<ToDoListItemButton
key={0}
onClick={onDeleteClick}
>
Delete
</ToDoListItemButton>
</ToDoListItem>
);
};
DeletableToDoListItem.propTypes = {
children: PropTypes.node,
id: PropTypes.number,
onDeleteClick: PropTypes.func,
};
Can you see what it's doing? It's going to take the children
we're already defining in our app
and add on additonal children, which is a ToDoListItemButton
in this case. The original
ToDoListItem
is fine with this, in fact it's not even aware that we're doing this. All it's
concerned with is accepting children and wrapping them inside of a div
. This flexibility is one
of the great advantages of composable components.
Back in our To-do app, we need to update the renderToDos
method to return instances of our new
DeletableToDoListItem
component instead of ToDoLisItem
. We also need to provide it with a
new id
prop and a new onDeleteClick
prop. Let's provide it with a reference to a method
called handleDeleteToDoClick
and define it like this:
handleDeleteToDoClick(toDoId) {
this.setState({
toDos: this.state.toDos.filter(toDo => toDo.id !== toDoId),
});
}
See how it accepts a reference to a To-do ID? See how this is the same ID we provided as a prop?
This pattern of providing components with props that they'll need to provide as arguments to
callbacks is recommened by AirBnB
as an alternative to creating a partially-applied function, e.g. this.handleDeleteToDoClick.bind(this, toDo.id)
.
Take-aways
- Leverage the composability of components to avoid making backwards-compatibility-breaking changes.
- Provide components with the data they'll need to provide to their callbacks.
Let's face it, we all procrastinate. Things pile up. Eventually we get up off our lazy ass and do one thing in the pile to feel better about ourselves or to get our significant other to just stop, I did it already, okay. But which one to do? Search to the rescue!
Up to this point we've been building UI first, and then writing logic later. Let's switch things up. Remember how we learned about the concept of derived state earlier? Well, this concept applies here.
We know the user will specify a search term, which we'll use to filter our To-dos, which will determine which ones get rendered.
From this series of events, you can see how the search term should live on the state
object, and the
filtering logic should be used to derive the filtered state. Here are the two methods we'll need to
add to our To-do app to make this happen:
searchToDos(searchTerm) {
this.setState({
searchTerm,
});
}
getRenderableToDos() {
function getFilteredToDos(searchTerm, toDos) {
return toDos.filter(toDo =>
toDo.body.trim().toLowerCase().indexOf(searchTerm.trim().toLowerCase()) !== -1
);
}
const searchTerm = this.state.searchTerm;
const renderableToDos =
searchTerm.trim().length
? getFilteredToDos(searchTerm, this.state.toDos)
: this.state.toDos.slice(0);
return renderableToDos;
}
Then, in the renderToDos
method we already have, just map over getRenderableToDos
instead of
mapping over state.toDos
directly.
Note: getRenderableToDos is a pure function. This makes it a good candidate for extraction into a general-use module in the future, so we can use the same search functionality everywhere in our app.
We can reuse our existing components to build the search UI:
<Label>
Search
</Label>
<form onSubmit={this.handleSearchFormSubmit}>
<TextInput
value={this.state.searchTerm}
onChange={this.handleSearchInputChange}
/>
</form>
Add the appropriate event handlers:
handleSearchFormSubmit(event) {
event.preventDefault();
this.searchToDos(this.state.searchTerm);
}
handleSearchInputChange(event) {
this.searchToDos(event.target.value);
}
Take-aways
- Procrastination is opportunity's assassin. (Victor Kiam)
- If you want to make an easy job seem mighty hard, just keep putting off doing it. (Olin Miller)
- The best way to get something done is to begin. (Unknown)
Some of the CSS for these exists, some of it does not. You're on your own with these! Good luck!
Your challenge is to change the appearance of a To-do, based on whether it's "critical" or not.
Hint: Add "isCritical" metadata to the ToDo model. Add a "className" prop to the ToDoListItem. When a To-do is critical, pass in a special class which alters the appearance of the To-do.
Your challenge is to allow users to change the body of an existing To-do.
Hint: Create a special component for editing To-dos, called something like ToDoEditor
.
Track the ID of the To-do that's being edited. When rendering the To-dos, render the ToDoEditor in
that To-do's place.
Your challenge is to alternate the sort direction when the user clicks the "Sort" button.
Hint: Continue in the same direction we established when we added Search functionality.