JotaiJotai

状態
Primitive and flexible state management for React

Effect

jotai-effect is a utility package for reactive side effects.

install

npm i jotai-effect

atomEffect

atomEffect is a utility function for declaring side effects and synchronizing atoms in Jotai. It is useful for observing and reacting to state changes.

Parameters

type CleanupFn = () => void
type EffectFn = (
get: Getter,
set: Setter & { recurse: Setter },
) => CleanupFn | void
function atomEffect(effectFn: EffectFn): Atom<void>

effectFn (required): A function for listening to state updates with get and writing state updates with set. The effectFn is useful for creating side effects that interact with other Jotai atoms. You can cleanup these side effects by returning a cleanup function.

Usage

Subscribe to Atom Changes

import { atomEffect } from 'jotai-effect'
const loggingEffect = atomEffect((get, set) => {
// runs on mount or whenever someAtom changes
const value = get(someAtom)
loggingService.setValue(value)
})

Setup and Teardown Side Effects

import { atomEffect } from 'jotai-effect'
const subscriptionEffect = atomEffect((get, set) => {
const unsubscribe = subscribe((value) => {
set(valueAtom, value)
})
return unsubscribe
})

Mounting with Atoms or Hooks

After defining an effect using atomEffect, it can be integrated within another atom's read function or passed to Jotai hooks.

const anAtom = atom((get) => {
// mounts the atomEffect when anAtom mounts
get(loggingEffect)
// ...
})
// mounts the atomEffect when the component mounts
function MyComponent() {
useAtom(subscriptionEffect)
// ...
}

The atomEffect behavior

  • Cleanup Function: The cleanup function is invoked on unmount or before re-evaluation.

    Example
    atomEffect((get, set) => {
    const intervalId = setInterval(() => set(clockAtom, Date.now()))
    return () => clearInterval(intervalId)
    })
  • Resistent To Infinite Loops: atomEffect does not rerun when it changes a value that it is watching with set.

    Example
    const countAtom = atom(0)
    atomEffect((get, set) => {
    // this will not infinite loop
    get(countAtom) // after mount, count will be 1
    set(countAtom, increment)
    })
  • Supports Recursion: Recursion is supported with set.recurse for both sync and async use cases, but is not supported in the cleanup function.

    Example
    const countAtom = atom(0)
    atomEffect((get, set) => {
    // increments count once per second
    const count = get(countAtom)
    const timeoutId = setTimeout(() => {
    set.recurse(countAtom, increment)
    }, 1000)
    return () => clearTimeout(timeoutId)
    })
  • Executes In The Next Microtask: effectFn runs in the next available microtask, after all Jotai synchronous read evaluations have completed.

    Example
    const countAtom = atom(0)
    const logAtom = atom([])
    const logCounts = atomEffect((get, set) => {
    set(logAtom, (curr) => [...curr, get(countAtom)])
    })
    const setCountAndReadLog = atom(null, async (get, set) => {
    get(logAtom) // [0]
    set(countAtom, increment) // effect runs in next microtask
    get(logAtom) // [0]
    await Promise.resolve().then()
    get(logAtom) // [0, 1]
    })
    store.set(setCountAndReadLog)
  • Batches Synchronous Updates (Atomic Transactions): Multiple synchronous updates to atomEffect atom dependencies are batched. The effect is run with the final values as a single atomic transaction.

    Example
    const enabledAtom = atom(false)
    const countAtom = atom(0)
    const updateEnabledAndCount = atom(null, (get, set) => {
    set(enabledAtom, (value) => !value)
    set(countAtom, (value) => value + 1)
    })
    const combos = atom([])
    const combosEffect = atomEffect((get, set) => {
    set(combos, (arr) => [...arr, [get(enabledAtom), get(countAtom)]])
    })
    store.set(updateEnabledAndCount)
    store.get(combos) // [[false, 0], [true, 1]]
  • Conditionally Running atomEffect: atomEffect is active only when it is mounted within the application. This prevents unnecessary computations and side effects when they are not needed. You can disable the effect by unmounting it.

    Example
    atom((get) => {
    if (get(isEnabledAtom)) {
    get(effectAtom)
    }
    })
  • Idempotent: atomEffect runs once when state changes regardless of how many times it is mounted.

    Example
    let i = 0
    const effectAtom = atomEffect(() => {
    get(countAtom)
    i++
    })
    const mountTwice = atom(() => {
    get(effectAtom)
    get(effectAtom)
    })
    store.set(countAtom, increment)
    Promise.resolve.then(() => {
    console.log(i) // 1
    })

Dependency Management

Aside from mount events, the effect runs when any of its dependencies change value.

  • Sync: All atoms accessed with get during the synchronous evaluation of the effect are added to the atom's internal dependency map.

    Example
    atomEffect((get, set) => {
    // updates whenever `anAtom` changes value but not when `anotherAtom` changes value
    get(anAtom)
    setTimeout(() => {
    get(anotherAtom)
    }, 5000)
    })
  • Async: For async effects, you should use an abort controller to cancel pending fetch requests and promises.

    Example
    atomEffect((get, set) => {
    const count = get(countAtom) // countAtom is an atom dependency
    const abortController = new AbortController()
    ;(async () => {
    try {
    await delay(1000)
    abortController.signal.throwIfAborted()
    get(dataAtom) // dataAtom is not an atom dependency
    } catch (e) {
    if (e instanceof AbortError) {
    // async cleanup logic here
    } else {
    console.error(e)
    }
    }
    })()
    return () => {
    // abort when countAtom changes
    abortController.abort(new AbortError())
    }
    })
  • Cleanup: Accessing atoms with get in the cleanup function does not add them to the atom's internal dependency map.

    Example
    atomEffect((get, set) => {
    // runs once on mount
    // does not update when `idAtom` changes
    const unsubscribe = subscribe((valueAtom) => {
    const value = get(valueAtom)
    // ...
    })
    return () => {
    const id = get(idAtom)
    unsubscribe(id)
    }
    })
  • Recalculation of Dependency Map: The dependency map is recalculated on every run. If an atom was not watched during the current run, it will not be in the current run's dependency map. Only actively watched atoms are considered dependencies.

    Example
    const isEnabledAtom = atom(true)
    atomEffect((get, set) => {
    // if `isEnabledAtom` is true, runs when `isEnabledAtom` or `anAtom` changes value
    // otherwise runs when `isEnabledAtom` or `anotherAtom` changes value
    if (get(isEnabledAtom)) {
    const aValue = get(anAtom)
    } else {
    const anotherValue = get(anotherAtom)
    }
    })

Comparison with useEffect

Component Side Effects

useEffect is a React Hook that lets you synchronize a component with an external system.

Hooks are functions that let you “hook into” React state and lifecycle features from function components. They are a way to reuse, but not centralize, stateful logic. Each call to a hook has a completely isolated state. This isolation can be referred to as component-scoped. For synchronizing component props and state with a Jotai atom, you should use the useEffect hook.

Global Side Effects

For setting up global side-effects, deciding between useEffect and atomEffect comes down to developer preference. Whether you prefer to build this logic directly into the component or build this logic into the Jotai state model depends on what mental model you adopt.

atomEffects are more appropriate for modeling behavior in atoms. They are scoped to the store context rather than the component. This guarantees that a single effect will be used regardless of how many calls they have.

The same guarantee can be achieved with the useEffect hook if you ensure that the useEffect is idempotent.

atomEffects are distinguished from useEffect in a few other ways. They can directly react to atom state changes, are resistent to infinite loops, and can be mounted conditionally.

It's up to you

Both useEffect and atomEffect have their own advantages and applications. Your project’s specific needs and your comfort level should guide your selection. Always lean towards an approach that gives you a smoother, more intuitive development experience. Happy coding!