This time I’ll give attention to the iterator design sample. The sample is closely used within the Swift normal library, there are protocols that gives you assist if you must create an iterator, however truthfully: I’ve by no means applied this sample instantly. 😅

The reality is that in all probability in 99% of the use circumstances you may by no means should take care of this sample, as a result of there may be superb assist for iterators built-in instantly into Swift. All the time use sequences, arrays, dictionaries as a substitute of instantly implementing this sample, but it surely’s good to know the way issues are working beneath the hood, is not it? 🙃

What’s the iterator design sample?

Because the identify suggests, the sample allows you to iterate over a group of components. Right here is the definition from the gang of 4 e book:

Present a strategy to entry the weather of an combination object sequentially with out exposing its underlying illustration.

Lengthy story brief the iterator provides you an interface that can allow you to iterate over collections no matter how they’re applied within the background. Here’s a fast instance of the idea above utilizing a string iterator.

import Basis

protocol StringIterator {
    func subsequent() -> String?
}

class ArrayStringIterator: StringIterator {

    personal let values: [String]
    personal var index: Int?

    init(_ values: [String]) {
        self.values = values
    }

    personal func nextIndex(for index: Int?) -> Int? {
        if let index = index, index < self.values.rely - 1 {
            return index + 1
        }
        if index == nil, !self.values.isEmpty {
            return 0
        }
        return nil
    }

    func subsequent() -> String? {
        if let index = self.nextIndex(for: self.index) {
            self.index = index
            return self.values[index]
        }
        return nil
    }
}


protocol Iterable {
    func makeIterator() -> StringIterator
}

class DataArray: Iterable {

    personal var dataSource: [String]

    init() {
        self.dataSource = ["🐶", "🐔", "🐵", "🦁", "🐯", "🐭", "🐱", "🐮", "🐷"]
    }

    func makeIterator() -> StringIterator {
        return ArrayStringIterator(self.dataSource)
    }
}

let knowledge = DataArray()
let iterator = knowledge.makeIterator()

whereas let subsequent = iterator.subsequent() {
    print(subsequent)
}

As you may see there are two principal protocols and a extremely easy implementation for each of them. Our DataArray class now acts like an actual array, the underlying components could be iterated via utilizing a loop. Let’s ditch the idea and re-implement the instance from above by utilizing actual Swift normal library elements. 😉

Customized sequences in Swift

Swift has a built-in sequence protocol that can assist you creating iterators. Implementing your personal sequence in Swift is all about hiding your underlying knowledge construction by making a customized iterator object. You simply should retailer the present index and return your subsequent ingredient in line with that every time the following operate will get known as. 😛

import Basis

struct Emojis: Sequence {
    let animals: [String]

    func makeIterator() -> EmojiIterator {
        return EmojiIterator(self.animals)
    }
}

struct EmojiIterator: IteratorProtocol {

    personal let values: [String]
    personal var index: Int?

    init(_ values: [String]) {
        self.values = values
    }

    personal func nextIndex(for index: Int?) -> Int? {
        if let index = index, index < self.values.rely - 1 {
            return index + 1
        }
        if index == nil, !self.values.isEmpty {
            return 0
        }
        return nil
    }

    mutating func subsequent() -> String? {
        if let index = self.nextIndex(for: self.index) {
            self.index = index
            return self.values[index]
        }
        return nil
    }
}

let emojis = Emojis(animals: ["🐶", "🐔", "🐵", "🦁", "🐯", "🐭", "🐱", "🐮", "🐷"])
for emoji in emojis {
    print(emoji)
}

So the Sequence protocol is a generic counterpart of our customized iterable protocol used within the first instance. The IteratorProtocol is considerably just like the string iterator protocol used earlier than, however extra Swift-ish and naturally extra generic.

So, that is nice. Lastly you know the way to create a customized sequence. Which is nice if you would like to cover your knowledge construction and supply a generic iterable interface. Think about what would occur should you had been about to start out utilizing a dictionary as a substitute of an array for storing named emojis with out an iterator that wraps them. 🤔

Now the factor is that there’s another tremendous helpful factor within the Swift normal library that I would like to speak about. That is proper, one abstraction stage up and right here we’re:

Customized collections in Swift

Collections are one step past sequences. Parts inside them could be accessed by way of subscript additionally they outline each a startIndex and an endIndex, plus particular person components of a group could be accessed a number of instances. Sounds good? 👍

Typically it may be helpful to create a customized assortment kind. For instance if you would like to get rid of elective values. Think about a categorized favourite mechanism, for each class you’d have an array of favorites, so that you’d should take care of empty and non-existing circumstances. With a customized assortment you may disguise that additional code inside your customized knowledge construction and supply a clear interface for the remainder of your app. 😍

class Favorites {

    typealias FavoriteType = [String: [String]]

    personal(set) var listing: FavoriteType

    public static let shared = Favorites()

    personal init() {
        self.listing = FavoriteType()
    }
}


extension Favorites: Assortment {

    typealias Index = FavoriteType.Index
    typealias Aspect = FavoriteType.Aspect

    var startIndex: Index {
        return self.listing.startIndex
    }
    var endIndex: Index {
        return self.listing.endIndex
    }

    subscript(index: Index) -> Iterator.Aspect {
        return self.listing[index]
    }

    func index(after i: Index) -> Index {
        return self.listing.index(after: i)
    }
}

extension Favorites {

    subscript(index: String) -> [String] {
        return self.listing[index] ?? []
    }

    func add(_ worth: String, class: String) {
        if var values = self.listing[category] {
            guard !values.accommodates(worth) else {
                return
            }
            values.append(worth)
            self.listing[category] = values
        }
        else {
            self.listing[category] = [value]
        }
    }

    func take away(_ worth: String, class: String) {
        guard var values = self.listing[category] else {
            return
        }
        values = values.filter { $0 == worth }

        if values.isEmpty {
            self.listing.removeValue(forKey: class)
        }
        else {
            self.listing[category] = values
        }
    }
}

Favorites.shared.add("apple", class: "fruits")
Favorites.shared.add("pear", class: "fruits")
Favorites.shared.add("apple", class: "fruits")

Favorites.shared["fruits"]

Favorites.shared.take away("apple", class: "fruits")
Favorites.shared.take away("pear", class: "fruits")
Favorites.shared.listing

I do know, it is a actually dumb instance, but it surely demonstrates why collections are extra superior in comparison with pure sequences. Additionally within the hyperlinks under there are nice demos of effectively written collections. Be happy to be taught extra about these tremendous protocols and customized knowledge varieties hidden (not so deep) contained in the Swift normal library. 🤐