Immer

Create the next immutable state tree by simply modifying the current tree

• NPM / Yarn: npm install immer
• CDN: https://unpkg.com/immer

Immer (German for: always) is a tiny package that allows you to work with immutable state in a more convenient way. It is based on the copy-on-write mechanism.

The basic idea is that you will apply all your changes to a temporarily draftState, which is a proxy of the currentState. Once all your mutations are completed, immer will produce the nextState based on the mutations to the draft state. This means that you can interact with your data by simply modifying it, while keeping all the benefits of immutable data.

Using immer is like having a personal assistant; he takes a letter (the current state), and gives you a copy (draft) to jot changes onto. Once you are done, the assistant will take your draft and produce the real immutable, final letter for you (the next state).

A mindful reader might notice that this is quite similar to withMutations of ImmutableJS. It is indeed, but generalized and applicable to plain, native JavaScript data structures (arrays and objects) without further needing any library.

Installation

npm install immer

API

The immer package exposes a single function:

immer(currentState, fn: (draftState) => void): nextState

Example

import immer from "immer"

const baseState = [
    {
        todo: "Learn typescript",
        done: true
    },
    {
        todo: "Try immer",
        done: false
    }
]

const nextState = immer(baseState, draftState => {
    draftState.push({ todo: "Tweet about it" })
    draftState[1].done = true
})

The interesting thing about immer is that, the baseState will be untouched, but the nextState will reflect all changes made to draftState.

// the new item is only added to the next state,
// base state is unmodified
expect(baseState.length).toBe(2)
expect(nextState.length).toBe(3)

// same for the changed 'done' prop
expect(baseState[1].done).toBe(false)
expect(nextState[1].done).toBe(true)

// unchanged data is structurally shared
expect(nextState[0]).toBe(baseState[0])
// changed data not (dûh)
expect(nextState[1]).not.toBe(baseState[1])

Using immer on older JavaScript environments

By default immer tries to use proxies for optimal performance. However, on older JavaScript engines Proxy is not available. For example, Microsoft Internet Explorer or React Native on Android. In these cases, import the ES5 compatibile implementation first, which is a bit slower (see below) but semantically equivalent:

import immer from "immer/es5"

Benefits

• Use the language© to construct create your next state
• Use JavaScript native arrays and object
• Automatic immutability; any state tree produced by immer will by defualt be deeply frozen
• Strongly typed, no string based paths etc
• Deep updates are trivial
• Small, dependency free library with minimal api surface
• No accidental mutations of current state, but intentional mutations of a draft state

Auto freezing

Immer automatically freezes any state trees that are modified using `immer. This protects against accidental modifications of the state tree outside of an immer function. This comes with a performance impact, so it is recommended to disable this option in production. It is by default enabled.

Use setAutoFreeze(true / false) to turn this feature on or off.

Reducer Example

Here is a simple example of the difference that this approach could make in practice. The todo reducers from the official Redux todos-with-undo example

Note, this is just a sample application of the immer package. Immer is not just designed to simplify Redux reducers. It can be used in any context where you have an immutable data tree that you want to clone and modify (with structural sharing)

const todo = (state, action) => {
  switch (action.type) {
    case 'ADD_TODO':
      return {
        id: action.id,
        text: action.text,
        completed: false
      }
    case 'TOGGLE_TODO':
      if (state.id !== action.id) {
        return state
      }

      return {
        ...state,
        completed: !state.completed
      }
    default:
      return state
  }
}

const todos = (state = [], action) => {
  switch (action.type) {
    case 'ADD_TODO':
      return [
        ...state,
        todo(undefined, action)
      ]
    case 'TOGGLE_TODO':
      return state.map(t =>
        todo(t, action)
      )
    default:
      return state
  }
}

After using immer, that simply becomes:

import immer from 'immer'

const todos = (state = [], action) =>
  // immer produces nextState from draftState and returns it
  immer(state, draftState => {
    switch (action.type) {
      case 'ADD_TODO':
        draftState.push({
          id: action.id,
          text: action.text,
          completed: false
        })
        return
      case 'TOGGLE_TODO':
        const todo = draftState.find(todo => todo.id === action.id)
        todo.completed = !todo.completed
        return
    }
  })

Creating middleware or a reducer wrapper that applies immer automatically is left as exercise for the reader :-).

Here are some typical reducer examples, take from the Redux Immutable Update Patterns page, and their immer counter part. These examples are semantically equivalent and produce the exact same state.

// Plain reducer
function insertItem(array, action) {
    return [
        ...array.slice(0, action.index),
        action.item,
        ...array.slice(action.index)
    ]
}

// With immer
function insertItem(array, action) {
    return immer(array, draft => {
        draft.splice(action.index, 0, action.item)
    })
}

// Plain reducer
function removeItem(array, action) {
    return [
        ...array.slice(0, action.index),
        ...array.slice(action.index + 1)
    ];
}

// With immer
function removeItem(array, action) {
    return immer(array, draft => {
        draft.splice(action.index, 1)
    })
}

// Plain reducer
function updateObjectInArray(array, action) {
    return array.map( (item, index) => {
        if(index !== action.index) {
            // This isn't the item we care about - keep it as-is
            return item;
        }

        // Otherwise, this is the one we want - return an updated value
        return {
            ...item,
            ...action.item
        };
    });
}

// With immer
function updateObjectInArray(array, action) {
    return immer(array, draft => {
        draft[action.index] = { ...item, ...action.item}
        // Alternatively, since arbitrarily deep updates are supported:
        // Object.assign(draft[action.index], action.item)
    })
}

Limitations

• Currently, only tree shaped states are supported. Cycles could potentially be supported as well (PR's welcome)
• Currently, only supports plain objects and arrays. Non-plain data structures (like Map, Set) not (yet). (PR's welcome)

Pitfalls:

• Immer only processes native arrays and plain objects (with a prototype of null or Object). Any other type of value will be treated verbatim! So if you modify a Map or Buffer or whatever complex object from the draft state, it will be that very same object in both the base state as the next state. In such cases, make sure to always produce fresh instances if you want to keep your state truly immutable.
• Make sure to modify the draft state you get passed in in the callback function, not the original current state that was passed as the first argument to immer!
• Since immer uses proxies, reading huge amounts of data from state comes with an overhead. If this ever becomes an issue (measure before you optimize!), do the current state analysis before entering the immer block or read from the currentState rather than the draftState
• Some debuggers (at least Node 6 is known) have trouble debugging when Proxies are in play. Node 8 is known to work correctly.

Performance

Here is a simple benchmark on the performance of immer. This test takes 100.000 todo items, and updates 10.000 of them. Freeze indicates that the state tree has been frozen after producing it. This is a development best practice, as it prevents developers from accidentally modifying the state tree.

These tests were executed on Node 8.4.0. Use yarn test:perf to reproduce them locally.

Some observations:

• The mutate, and deepclone, mutate benchmarks establish a baseline on how expensive changing the data is, without immutability (or structural sharing in the deep clone case).
• The reducure and naive reducer are implemented in typical Redux style reducers. The "smart" implementation slices the collection first, and then maps and freezes only the relevant todos. The "naive" implementation just maps over and processes the entire collection.
• Immer with proxies is roughly speaking twice as slow as a hand written reducer. This is in practice negclectable.
• Immer is roughly as fast as ImmutableJS. However, the immutableJS + toJS makes clear the cost that often needs to be paid later; converting the immutableJS objects back to plain objects, to be able to pass them to components, over the network etc... (And there is also the upfront cost of converting data received from e.g. the server to immutable JS)
• The ES5 implentation of immer is significantly slower. For most reducers this won't matter, but reducers that process large amounts of data might benefit from not (or only partially) using an immer producer. Luckily, immer is fully opt-in.
• The peeks in the frozen versions of just mutate, deepclone and naive reducer come from the fact that they recursively freeze the full state tree, while the other test cases only freeze the modified parts of the tree.

Credits

Special thanks goes to @Mendix, which supports it's employees to experiment completely freely two full days a month, which formed the kick-start for this project.