This package provides a lightweight and flexible solution for rendering large tree structures. It is built on top of the react-window library.
Attention! This library is entirely rewritten to work with the react-window. If you are looking for the tree view solution for the react-virtualized, take a look at react-virtualized-tree.
NOTE: This is the documentation for version 3.x.x. For version 2.x.x see this branch.
# npm
npm i react-window react-vtree
# Yarn
yarn add react-window react-vtreeYou can also take a look at the very similar example at the Storybook:
import {FixedSizeTree as Tree} from 'react-vtree';
// Tree component can work with any possible tree structure because it uses an
// iterator function that the user provides. Structure, approach, and iterator
// function below is just one of many possible variants.
const treeNodes = [
{
name: 'Root #1',
id: 'root-1',
children: [
{
children: [
{id: 'child-2', name: 'Child #2'},
{id: 'child-3', name: 'Child #3'},
],
id: 'child-1',
name: 'Child #1',
},
{
children: [{id: 'child-5', name: 'Child #5'}],
id: 'child-4',
name: 'Child #4',
},
],
},
{
name: 'Root #2',
id: 'root-2',
},
];
// This helper function constructs the object that will be sent back at the step
// [2] during the treeWalker function work. Except for the mandatory `data`
// field you can put any additional data here.
const getNodeData = (node, nestingLevel) => ({
data: {
id: node.id.toString(), // mandatory
isLeaf: node.children.length === 0,
isOpenByDefault: true, // mandatory
name: node.name,
nestingLevel,
},
nestingLevel,
node,
});
// The `treeWalker` function runs only on tree re-build which is performed
// whenever the `treeWalker` prop is changed.
function* treeWalker() {
// Step [1]: Define the root node of our tree. There can be one or
// multiple nodes.
for (let i = 0; i < treeNodes.length; i++) {
yield getNodeData(treeNodes[i], 0);
}
while (true) {
// Step [2]: Get the parent component back. It will be the object
// the `getNodeData` function constructed, so you can read any data from it.
const parent = yield;
for (let i = 0; i < parent.node.children.length; i++) {
// Step [3]: Yielding all the children of the provided component. Then we
// will return for the step [2] with the first children.
yield getNodeData(parent.node.children[i], parent.nestingLevel + 1);
}
}
}
// Node component receives all the data we created in the `treeWalker` +
// internal openness state (`isOpen`), function to change internal openness
// state (`setOpen`) and `style` parameter that should be added to the root div.
const Node = ({data: {isLeaf, name}, isOpen, style, setOpen}) => (
<div style={style}>
{!isLeaf && (
<button type="button" onClick={() => setOpen(!isOpen)}>
{isOpen ? '-' : '+'}
</button>
)}
<div>{name}</div>
</div>
);
ReactDOM.render(
<Tree treeWalker={treeWalker} itemSize={30} height={150} width={300}>
{Node}
</Tree>,
document.querySelector('#root'),
);You can read more about these properties in the FixedSizeList documentation.
children: componentclassName: string = ""direction: strig = "ltr"height: strig | numberinitialScrollOffset: number = 0innerRef: function | createRef object. This property works as it described in thereact-window. For getting aFixedSizeListreference uselistRef.innerElementType: React.ElementType = "div"innerTagName: stringreact-window.itemData: anyitemKey: functionitemSize: numberlayout: string = "vertical"onItemsRendered: functiononScroll: functionouterRef: function | createRef objectouterElementType: React.ElementType = "div"outerTagName: stringreact-window.overscanCount: number = 1style: object = nulluseIsScrolling: boolean = falsewidth: number | string
This option allows making the tree asynchronous; e.g. you will be able to load the branch data on the node opening. All it does under the hood is preserving the tree state between tree buildings on treeWalker update, so the user does not see the tree resetting to the default state when the async action is performed.
To see how it works you can check the AsyncData story. You can use the disableAsync to see what will happen on the async action if the async prop is false.
If it is combined with the placeholder option, the tree re-building won't be interrupted by showing the placeholder; it will be shown only at the first time the tree is building.
To see how two options interact with each other see the AsyncDataIdle story.
The Node component responsible for rendering each node.
It receives the following props:
-
Inherited from
react-window'sRowcomponent:style: objectisScrolling: boolean- ifuseIsScrollingis enabled.
-
Node-specific props:- All fields of the
FixedSizeNodePublicStateobject. treeData: any- any data provided via theitemDataproperty of theFixedSizeTreecomponent.
- All fields of the
This property receives any react node that will be displayed instead of a tree during the building process. This option should only be used if the tree building process requires too much time (which means you have a really giant amount of data, e.g. about a million nodes).
Setting this option enables the requestIdleCallback under the hood for browsers that support this feature. For other browsers the original scenario is applied; no placeholder will be shown.
Using this feature allows avoiding UI freezes; however, it may slightly increase the time spent for the building process.
To see how it works, you can check the BigData story. Use placeholder tool to add and remove placeholder.
If you have an asynchronous giant tree and want to use profits of requestIdleCallback but don't want placeholder to be shown on the first render (that is probably quite small because all other data will be loaded asynchronously), set placeholder to null. No placeholder will be shown on the first render but the requestIdleCallback building will be enabled and allow avoiding freezes on tree re-building when tree becomes bigger.
To see how it works you can check the AsyncDataIdle story. It uses the null placeholder, so no text is shown for the first build but async requests don't block the UI.
This option works in tandem with the placeholder option. With it, you can set the task timeout for the requestIdleCallback. The buildingTaskTimeout will be sent directly as the requestIdleCallback's timeout option.
This option allows you to get the instance of the internal react-window list. It is usually unnecessary because all necessary methods are already provided but still can be useful for edge cases.
This property receives a custom Row component for the FixedSizeList that will override the default one. It can be used for adding new functionality.
Row component receives the following props:
index: numberdata: object- the data tree component provides toRow. It contains the following data:component: component- aNodecomponent to create a React element from.getRecordData: function- a function that gets the record data byindex. It returns aFixedSizeNodePublicStateobject.treeData: any- any data provided via theitemDataproperty of theFixedSizeTreecomponent.
style: objectisScrolling: boolean
An iterator function that walks around the tree and yields node data to build an inner representation of the tree. For algorithm details, see TreeWalker section.
The treeWalker function should yield the object of the following shape:
data: FixedSizeNodeData- this field is mandatory. SeeFixedSizeNodeDatatype for the shape....- you can add any other data to this object. It will be sent directly to thetreeWalkerat the step [2] of the execution.
Tree is re-computed on each treeWalker change. To avoid unnecessary tree re-computation keep the treeWalker memoized (e.g. with useCallback hook). If you want to update tree data, send the new version of treeWalker to the tree component.
Note that when treeWalker is updated no internal state will be shared with the new tree. Everything will be built from scratch.
The component provides the following methods.
The scrollToItem method behaves the same as scrollToItem from FixedSizeList but receives node id instead of index.
This method starts the tree traversing to update the internal state of nodes.
It receives state object that contains nodes' id as keys and update rules as values. Each record traverses a subtree of the specified node (also "owner node") and does not affect other nodes (it also means that if you specify the root node the whole tree will be traversed).
The rules object has the following shape:
open: boolean- this rule changes the openness state for the owner node only (subtree nodes are not affected).subtreeCallback(node: object, ownerNode: object): void- this callback runs against each node in the subtree of the owner node (including the owner node as well). It receives the subtree node and the owner node. Changing any property of the subtree node will affect the node state and how it will be displayed (e.g. if you change the node openness state it will be displayed according to the changed state).
The order of rules matters. If you specify the child node rules before the parent node rules, and that rules affect the same property, the parent node subtreeCallback will override that property. So if you want to override parent's rules, place children rules after the parent's.
The type of the node objects received by subtreeCallback is FixedSizeNodePublicState. See the types description below.
// The tree
const tree = {
name: 'Root #1',
id: 'root-1',
children: [
{
children: [
{id: 'child-2', name: 'Child #2'},
{id: 'child-3', name: 'Child #3'},
],
id: 'child-1',
name: 'Child #1',
},
{
children: [{id: 'child-5', name: 'Child #5'}],
id: 'child-4',
name: 'Child #4',
},
],
};
// recomputeTree
tree.recomputeTree({
'root-1': {
open: false,
subtreeCallback(node, ownerNode) {
// Since subtreeCallback affects the ownerNode as well, we can check if the
// nodes are the same, and run the action only if they aren't
if (node !== ownerNode) {
// All nodes of the tree will be closed
node.isOpen = false;
}
},
},
// But we want `child-4` to be open
'child-4': true,
});FixedSizeNodeData- value of thedatafield of the object yielded by thetreeWalkerfunction. The shape is the following:id- a unique identifier of the node.isOpenByDefault- a default openness state of the node....- you can add any number of additional fields to this object. This object without any change will be sent directly to theNodecomponent. You can also usegetRecordDatafunction to get this object along with the other record data by the index. To describe that data, you have to create a new type that extends theFixedSizeNodeDatatype.
FixedSizeNodePublicState<TData extends FixedSizeNodeData>- the node state available for theNodecomponent andrecomputeTree'ssubtreeCallbackfunction. It has the following shape:data: FixedSizeNodeData.isOpen: boolean- a current openness status of the node.setOpen(state: boolean): function- a function to change the openness state of the node. It receives the new openness state as abooleanand opens/closes the node accordingly.
FixedSizeTreeProps<TData extends FixedSizeNodeData>- props thatFixedSizeTreecomponent receives. Described in the Props section.FixedSizeTreeState<TData extends FixedSizeNodeData>- state thatFixedSizeTreecomponent has.
You can also take a look at the very similar example at the Storybook:
import {VariableSizeTree as Tree} from 'react-vtree';
// Tree component can work with any possible tree structure because it uses an
// iterator function that the user provides. Structure, approach, and iterator
// function below is just one of many possible variants.
const tree = {
name: 'Root #1',
id: 'root-1',
children: [
{
children: [
{id: 'child-2', name: 'Child #2'},
{id: 'child-3', name: 'Child #3'},
],
id: 'child-1',
name: 'Child #1',
},
{
children: [{id: 'child-5', name: 'Child #5'}],
id: 'child-4',
name: 'Child #4',
},
],
};
// This helper function constructs the object that will be sent back at the step
// [2] during the treeWalker function work. Except for the mandatory `data`
// field you can put any additional data here.
const getNodeData = (node, nestingLevel) => ({
data: {
defaultHeight: itemSize, // mandatory
id: node.id.toString(), // mandatory
isLeaf: node.children.length === 0,
isOpenByDefault: true, // mandatory
name: node.name,
nestingLevel,
},
nestingLevel,
node,
});
// The `treeWalker` function runs only on tree re-build which is performed
// whenever the `treeWalker` prop is changed.
function* treeWalker() {
// Step [1]: Define the root node of our tree. There can be yielded one or
// multiple nodes.
yield getNodeData(tree, 0);
while (true) {
// Step [2]: Get the parent component back. It will be the object
// the `getNodeData` function constructed, so you can read any data from it.
const parent = yield;
for (let i = 0; i < parent.node.children.length; i++) {
// Step [3]: Yielding all the children of the provided component. Then we
// will return for the step [2] with the first children.
yield getNodeData(parent.node.children[i], parent.nestingLevel + 1);
}
}
}
// Node component receives current node height as a prop
const Node = ({data: {isLeaf, name}, height, isOpen, style, setOpen}) => (
<div style={style}>
{!isLeaf && (
<button type="button" onClick={() => setOpen(!isOpen)}>
{isOpen ? '-' : '+'}
</button>
)}
<div>{name}</div>
</div>
);
const Example = () => (
<Tree treeWalker={treeWalker} height={150} width={300}>
{Node}
</Tree>
);You can read more about these properties in the VariableSizeList documentation.
Since VariableSizeList in general inherits properties from the FixedSizeList, everything described in the same section for FixedSizeTree affects this section. For the rest, there are the following changes:
estimatedItemSize: number = 50itemSize: (index: number) => number. This property is optional. If it is not provided, thedefaultHeightof the specific node will be used. Advanced property; prefer using node state for it.
The Node component. It is the same as the FixedSizeTree's one but receives properties from the VariableSizeNodePublicState object.
Same as listRef of FixedSizeTree.
See rowComponent in the FixedSizeTree section; the getRecordData returns the VirtualSizeNodePublicState object.
An iterator function that walks over the tree. It behaves the same as FixedSizeTree's treeWalker. The data object should be in the VariableSizeNodeData shape.
The component provides the following methods:
The scrollToItem method behaves the same as scrollToItem from VariableSizeList but receives node id instead of index.
This method replaces the resetAfterIndex method of VariableSizeList but works exactly the same. It receives node id as a first argument.
See FixedSizeTree's recomputeTree description. There are no differences.
All types in this section are the extended variants of FixedSizeTree types.
VariableSizeNodeData- this object extendsFixedSizeNodeDataand contains the following additional fields:defaultHeight: number- the default height the node will have.
VariableSizeNodePublicState<TData extends VariableSizeNodeData>. The node state object. Extends theFixedSizeNodePublicStateand contains the following additional fields:height: number- the current height of the node. The node will be displayed with this height.resize(newHeight: number, shouldForceUpdate?: boolean): function- a function to change the height of the node. It receives two parameters:newHeight: number- a new height of the node.shouldForceUpdate: boolean- an optional argument that will be sent directly to theresetAfterIndexmethod.
VariableSizeTreeProps<T extends VariableSizeNodeData>.VariableSizeTreeState<T extends VariableSizeNodeData>.
The treeWalker algorithm works in the following way. During the execution, the treeWalker function sends a bunch of objects to the tree component which builds an internal representation of the tree. However, for it, the specific order of yieldings should be performed.
- The first yielding is always root nodes. They will be the foundation of the whole tree.
- Now start a loop where you will receive the parent node and yield all the children of it.
- The first yielding of loop iteration should yield an
undefined. In exchange, you will receive a node for which you should yield all the children in the same way you've done with the root ones. - When all the children are yielded, and the new iteration of loop is started, you yield
undefinedagain and in exchange receive the next node. It may be:- a child node if the previous node has children;
- a sibling node if it has siblings;
- a sibling of the elder node.
- When the whole tree is finished and algorithm reaches the end, the loop stops. You don't have to finish
treeWalker's loop manually.
The example of this algorithm is the following treeWalker function:
function* treeWalker() {
// Here we start our tree by yielding the data for the root node.
yield getNodeData(rootNode, 0);
while (true) {
// Here in the loop we receive the next node whose children should be
// yielded next.
const parent = yield;
for (let i = 0; i < parent.node.children.length; i++) {
// Here we go through the parent's children and yield them to the tree
// component
yield getNodeData(parent.node.children[i], parent.nestingLevel + 1);
// Then the loop iteration is over, and we are going to our next parent
// node.
}
}
}If you use react-vtree of version 2, it is preferable migrate to the version 3. The third version is quite different under the hood and provides way more optimized approach to the initial tree building and tree openness state change. The most obvious it becomes if you have a giant tree (with about 1 million of nodes).
To migrate to the new version, you have to do the following steps.
The treeWalker was and is the heart of the react-vtree. However, now it looks a bit different.
Old treeWalker worked for both initial tree building and changing node openness state:
function* treeWalker(refresh) {
const stack = [];
stack.push({
nestingLevel: 0,
node: rootNode,
});
// Go through all the nodes adding children to the stack and removing them
// when they are processed.
while (stack.length !== 0) {
const {node, nestingLevel} = stack.pop();
const id = node.id.toString();
// Receive the openness state of the node we are working with
const isOpened = yield refresh
? {
id,
isLeaf: node.children.length === 0,
isOpenByDefault: true,
name: node.name,
nestingLevel,
}
: id;
if (node.children.length !== 0 && isOpened) {
for (let i = node.children.length - 1; i >= 0; i--) {
stack.push({
nestingLevel: nestingLevel + 1,
node: node.children[i],
});
}
}
}
}The new treeWalker is only for the tree building. The Tree component builds and preserves the tree structure internally. See the full description above.
// This function prepares an object for yielding. We can yield an object
// that has `data` object with `id` and `isOpenByDefault` fields.
// We can also add any other data here.
const getNodeData = (node, nestingLevel) => ({
data: {
id: node.id.toString(),
isLeaf: node.children.length === 0,
isOpenByDefault: true,
name: node.name,
nestingLevel,
},
nestingLevel,
node,
});
function* treeWalker() {
// Here we send root nodes to the component.
for (let i = 0; i < rootNodes.length; i++) {
yield getNodeData(rootNodes[i], 0);
}
while (true) {
// Here we receive an object we created via getNodeData function
// and yielded before. All we need here is to describe its children
// in the same way we described the root nodes.
const parentMeta = yield;
for (let i = 0; i < parentMeta.node.children.length; i++) {
yield getNodeData(
parentMeta.node.children[i],
parentMeta.nestingLevel + 1,
);
}
}
}Components haven't been changed a lot but you may want to add new features like:
The recomputeTree method now receives a list of nodes to change (previously, it was an opennessState object). See the full description above.
The most important change is the introduction of the subtreeCallback. It is a function that will be applied to each node in the subtree of the specified node. Among other useful things it also allows imitating the behavior of old useDefaultOpenness and useDefaultHeight options.
Old recomputeTree:
treeInstance.recomputeTree({
opennessState: {
'node-1': true,
'node-2': true,
'node-3': false,
},
refreshNodes: true,
useDefaultOpenness: false,
});New recomputeTree:
treeInstance.recomputeTree({
'node-1': true,
'node-2': {
open: true,
subtreeCallback(node, ownerNode) {
if (node !== ownerNode) {
node.isOpen = false;
}
},
},
'node-3': false,
});In the 3.x.x version node provides a setOpen function instead of toggle that allows more fine-grained control over the openness state.
Old toggle:
const Node = ({data: {isLeaf, name}, isOpen, style, toggle}) => (
<div style={style}>
{!isLeaf && (
<div>
<button onClick={toggle}>{isOpen ? '-' : '+'}</button>
</div>
)}
<div>{name}</div>
</div>
);New setOpen:
const Node = ({data: {isLeaf, name}, isOpen, style, setOpen}) => (
<div style={style}>
{!isLeaf && (
<div>
// Imitating the old `toggle` function behavior
<button onClick={() => setOpen(!isOpen)}>{isOpen ? '-' : '+'}</button>
</div>
)}
<div>{name}</div>
</div>
);Using node IDs as keys should improve React rendering performance. However, it means that you won't be able to use Symbol as IDs anymore. You should move all your IDs to be strings instead of symbols.
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent