🟣 Verge is a very tunable state-management engine on iOS App (UIKit / SwiftUI) and built-in ORM.
đź“ŤAn effective state management architecture for iOS - UIKit, SwiftUIđź“Ť
_ An easier way to get unidirectional data flow _
_ Supports concurrent processing _
Verge is a high-performance, scalable state management library for Swift, designed with real-world use cases in mind. It offers a lightweight and easy-to-use approach to managing your application state without the need for complex actions and reducers. This guide will walk you through the basics of using Verge in your Swift projects.
Verge was designed with the following concepts in mind:
commit
function that accepts a closure describing how to change the state.To use Verge, follow these steps:
Equatable
protocol.Store
with your initial state.commit
method on the store instance.sinkState
method.Create a state struct that represents the state of your application. Your state struct should conform to the Equatable
protocol. This allows Verge to detect changes in your state and trigger updates as necessary.
Example:
struct MyState: StateType {
var count: Int = 0
}
Create a Store
instance with the initial state of your application. The Store
class takes two type parameters:
Never
.Example:
let store = Store<_, Never>(initialState: MyState())
To update your application state, use the commit
method on your Store
instance. The commit
method takes a closure with a single parameter, which is a mutable reference to your state. Inside the closure, modify the state as needed.
Example:
store.commit {
$0.count += 1
}
To receive updates when the state changes, use the sinkState
method on your Store
instance. This method takes a closure that receives the updated state as its parameter. The closure will be called whenever the state changes.
Example:
store.sinkState { state in
// Receives updates of the state
}
.storeWhileSourceActive()
The storeWhileSourceActive()
call at the end is a method provided by Verge to automatically manage the lifetime of the subscription. It retains the subscription as long as the source (in this case, the store
instance) is alive.
That’s it! You now know the basics of using Verge to manage the state in your Swift applications. For more advanced use cases and additional features, please refer to the official Verge documentation and GitHub repository.
In certain scenarios, event-driven programming is essential for creating responsive and efficient applications. The Verge library’s Activity of Store feature is designed to cater to this need, allowing developers to handle events seamlessly within their projects.
The Activity of Store comes into play when your application requires event-driven programming. It enables you to manage events and associated logic independently from the main store management, promoting a clean and organized code structure. This separation of concerns simplifies the overall development process and makes it easier to maintain and extend your application over time.
By leveraging the Activity of Store functionality, you can efficiently handle events within your application while keeping the store management intact. This ensures that your application remains performant and scalable, enabling you to build robust and reliable Swift applications using the Verge library.
Here’s an example of using Activity of Store:
let store: Store<MyState, MyActivity>
store.send(MyActivity.somethingHappened)
store.sinkActivity { (activity: MyActivity) in
// handle activities.
}
.storeWhileSourceActive()
To use Verge in SwiftUI, you can utilize the StoreReader
to subscribe to state updates within your SwiftUI views. Here’s an example of how to do this:
import SwiftUI
import Verge
struct ContentView: View {
@StoreObject private var viewModel = CounterViewModel()
var body: some View {
VStack {
StoreReader(viewModel.store) { state in
Text("Count: \(state.count)")
.font(.largeTitle)
}
Button(action: {
viewModel.increment()
}) {
Text("Increment")
}
}
}
}
final class CounterViewModel: StoreComponentType {
struct State: Equatable {
var count: Int = 0
}
let store: Store<State, Never> = .init(initialState: .init())
func increment() {
commit {
$0.count += 1
}
}
}
In this example, StoreReader
is used to read the state from the MyViewModel
store. This allows you to access and display the state within your SwiftUI view. Additionally, you can perform actions by calling methods on the store directly, as demonstrated with the button in the example.
This new section will help users understand how to use Verge with SwiftUI, allowing them to manage state effectively within their SwiftUI views. Let me know if you have any further suggestions or changes!
StoreObject property wrapper:
SwiftUI provides the @StateObject
property wrapper to create and manage a persistent instance of a given object that adheres to the ObservableObject protocol. However, StateObject will cause the view to be refreshed whenever the ObservableObject is updated.
In Verge, we introduce the StoreObject property wrapper, which instantiates a Store object for the duration of the view’s lifecycle but does not cause the view to refresh when the Store updates.
This is beneficial when you want to manage the Store in a more granular way, without causing the entire view to refresh when the Store changes. Instead, Store updates can be handled through the StoreReader.
Here’s a simple usage example of Verge with a UIViewController:
class MyViewController: UIViewController {
private struct State: Equatable {
var count: Int = 0
}
private let store: Store<State, Never> = .init(initialState: .init())
private let label: UILabel = .init()
override func viewDidLoad() {
super.viewDidLoad()
setupUI()
// Subscribe to the store's state updates
store.sinkState { [weak self] state in
guard let self = self else { return }
// Check if the value has been updated using ifChanged
state.ifChanged(\.count) { count in
self.label.text = "Count: \(count)"
}
}
.storeWhileSourceActive()
}
private func setupUI() {
// Omitted for brevity
}
private func incrementCount() {
store.commit {
$0.count += 1
}
}
}
sinkState
, Changed<State>
, and ifChanged
In UIKit, which is event-driven, it’s crucial to update components efficiently by only updating them as needed. The Verge library provides a way to achieve this using the sinkState
method, the Changed<State>
type, and the ifChanged
method.
When you use the sinkState
method, the closure you provide receives the latest state wrapped in a Changed<State>
type. This wrapper also includes the previous state, allowing you to determine which properties have been updated using the ifChanged
method.
Here’s an example of using sinkState
and ifChanged
in UIKit to efficiently update components:
store.sinkState {
$0.ifChanged(\.myProperty) { newValue in
// Update the component only when myProperty has changed
}
}
In this example, the component is updated only when myProperty
has changed, ensuring efficient updates in the UIKit-based application.
Compared to UIKit, SwiftUI works with a declarative view structure, which means that there is less need to check for state changes to update the view. However, when working with UIKit, using sinkState
, Changed<State>
, and ifChanged
helps maintain a performant and responsive application.
Verge’s Store includes a TaskManager that allows you to dispatch and manage asynchronous operations. This feature simplifies handling async tasks while keeping them associated with your Store.
To use TaskManager, simply call the task
method on your Store instance, and provide a closure that contains the asynchronous operation:
store.task {
await runMyOperation()
}
TaskManager also enables you to manage tasks based on keys and modes. You can assign a unique key to each task and specify a mode for its execution. This allows you to control the execution behavior of tasks based on their keys.
For example, you can use the .dropCurrent
mode to drop any currently running tasks with the same key and run the new task immediately:
store.task(key: .init("MyOperation"), mode: .dropCurrent) {
//
}
This functionality provides you with fine-grained control over how tasks are executed, ensuring that your application remains responsive and efficient, even when handling multiple asynchronous operations.
In theory, managing your entire application state in a single store is ideal. However, in large and complex applications, the computational complexity can become significant, leading to performance issues and slow application responsiveness. In such cases, it’s recommended to separate your state into multiple stores and integrate them as needed.
By dividing your state into multiple stores, you can reduce the complexity and overhead associated with state updates. Each store can manage a specific part of your application state, ensuring that updates are performed efficiently and quickly. This approach also promotes better organization and separation of concerns in your code, making it easier to maintain and extend your application over time.
To use multiple stores, create separate Store instances for different parts of your application state, and then connect them as needed. This may involve passing store instances to child components or sharing stores between sibling components. By structuring your application this way, you can ensure that each part of your application state is managed efficiently and effectively.
To copy state between stores, you can use the sinkState
method along with the ifChanged
function to only trigger updates when the state has changed. Here’s an example:
store.sinkState {
$0.ifChanged(\.myState) { value in
otherStore.commit {
$0.myState = value
}
}
}
In this example, when the state of myState
changes in store
, the new value is committed to otherStore
. This approach allows you to synchronize state between multiple stores efficiently.
Verge’s Derived
feature allows you to create computed properties based on your store’s state and efficiently subscribe to updates. This feature can help you optimize your application by reducing unnecessary computations and updates. Derived is inspired by the reselect library and provides similar functionality.
To create a derived property, you’ll use the store.derived
method. This method takes a Pipeline
object that describes how the derived data is generated:
let derived: Derived<Int> = store.derived(.select(\\.count))
You can use select
or map
to generate derived data. select
is used to take a value directly from the state, while map
can be used to generate new values based on the state, similar to a map function:
let derived: Derived<Int> = store.derived(.map { $0.count * 2 })
The Pipeline
checks if the derived data has been updated from the previous value. If it hasn’t changed, Derived
won’t publish any changes.
You can create another Derived instance from an existing Derived instance, effectively chaining them together:
let anotherDerived: Derived<String> = derived.derived(.map { $0.description })
To subscribe to updates of a derived property, you can use the sinkState
method, just like with a store:
derived.sinkState { value in
// Handle updates of the derived property
}
.storeWhileSourceActive()
By using Derived
for computed properties and subscribing to updates, you can ensure that your application remains efficient and performant, avoiding unnecessary computations and state updates.
State management plays a crucial role in building efficient and maintainable applications. One of the essential aspects of state management is organizing the data in a way that simplifies its manipulation and usage. This is where normalization becomes vital.
Normalization is the process of structuring data in a way that eliminates redundancy and ensures data consistency. It is essential in state-management libraries because it significantly reduces the computational complexity of operations and makes it easier to manage the state.
Docs:
Let’s take a look at an example to illustrate the difference between normalized and denormalized data structures.
Denormalized data structure:
posts:
- id: 1
title: "Post 1"
author:
id: 1
name: "Alice"
- id: 2
title: "Post 2"
author:
id: 1
name: "Alice"
- id: 3
title: "Post 3"
author:
id: 2
name: "Bob"
In the denormalized structure, author data is duplicated in each post, which can lead to inconsistencies and make it harder to manage the state.
Normalized data structure:
entities:
authors:
1:
id: 1
name: "Alice"
2:
id: 2
name: "Bob"
posts:
1:
id: 1
title: "Post 1"
authorId: 1
2:
id: 2
title: "Post 2"
authorId: 1
3:
id: 3
title: "Post 3"
authorId: 2
In the normalized structure, author data is stored separately from posts, eliminating data redundancy and ensuring data consistency. The relationship between posts and authors is represented by the authorId
field in the posts.
VergeORM is designed to handle normalization in state-management libraries effectively. By leveraging VergeORM, you can simplify your state management, reduce the computational complexity of operations, and improve the overall performance and maintainability of your application.
Defining Entities
Here’s an example of how to define the Book
and Author
entities:
struct Book: EntityType {
typealias EntityIDRawType = String
var entityID: EntityID {
.init(rawID)
}
let rawID: String
var name: String = "initial"
let authorID: Author.EntityID
}
struct Author: EntityType {
typealias EntityIDRawType = String
var entityID: EntityID {
.init(rawID)
}
let rawID: String
var name: String = ""
}
Defining Database Schema
To store the entities in the state, you need to define the database schema:
@NormalizedStorage
struct Database {
@Table
var books: Tables.Hash<Book> = .init()
@Table
var authors: Tables.Hash<Book> = .init()
}
Embedding the Database in State
Embed the Database
in your application’s state:
struct RootState: StateType {
var database: Database = .init()
}
Storing and Querying Entities
Here’s an example of how to store and query entities using a store
property
// Storing entities
store.commit {
$0.database.performBatchUpdates { context in
let authors = (0..<10).map { i in
Author(rawID: "\(i)")
}
let result = context.modifying.author.insert(authors)
}
}
// Querying entities
let book = store.state.database.db.book.find(by: .init("1"))
let author = store.state.database.db.author.find(by: .init("1"))
In this example, we use store.commit
to perform batch updates on the database. We insert a new set of authors into the author
entity table. Then, we use store.state.database.db
to query the book
and author
entities by their identifiers.
By using VergeNormalization, you can efficiently manage your application state with a normalized data structure, which simplifies your state management, reduces the computational complexity of operations, and improves the overall performance and maintainability of your application.
Verge supports SwiftPM.
This repo has several demo applications in Demo directory.
And we’re looking for your demo applications to list it here!
Please tell us from Issue!
🇯🇵 Muukii (Hiroshi Kimura)
Verge is released under the MIT license.