#Entities

You're viewing documentation for a previous version of this software. Switch to the latest stable version

Continuing from the setup, now we will be declaring entities for the GraphQL API.

Let's say for this example, we will have a user management system where the API can be used to create, find, update, and delete user profile(s).

#User profiles

Declaring the User model is fairly straight forward.

User.swift
struct User: Identifiable, Codable, Sendable {
    var id: UUID
    var username: String?
    var email: String
    var bio: String

    var displayName: String {
        username ?? email
    }

    var friendIDs: [UUID]
}

Here, we have a couple properties and some computed ones as well. All things should be pretty self-explanatory.

ID type

Graphiti cannot diffentiate String type from ID type by default. Pioneer brought in a custom ID struct to tackle this issue.

This customID can be constructed from any string with its regular initializer or from string literal(s). It can also come from UUID and String using the extension .toID() method.

_{It is also hashable so it can also be used as id requirement for Identifiable}

import Pioneer

extension User {
    var _id: ID {
        id.toID()
    }

    var _friendIDs: [ID] {
        friendIDs.map { $0.toID() }
    }
}

let schema = try Schema<Void, Resolver> {
    // Add as Scalar type, so Graphiti won't get mad
    ID.asScalar()
    // or
    Scalar(ID.self, as: "ID")
}

#User inputs

We also want an entity for the input that the API client will give to create and/or update a User profile.

import Pioneer

struct UserInput: Codable {
    var username: String?
    var email: String
    var bio: String
    var friendIDs: [ID]
}

extension User {
    init(id: ID = .uuid(), _ input: UserInput) {
        self.init(
            id: UUID(id.string) ?? UUID(),
            username: input.username,
            email: input.email,
            bio: input.bio,
            friendIDs: input.friendIDs.compactMap { UUID($0) },
        )
    }
}

#Datastore

In a real application, you want these User(s) to be stored in a persistent database like PostgreSQL or something similar. For this example, we will be simplying the workflow by just building one in memory.

Fluent FAQ

../../advanced/fluent/

We will be taking advantange of Swift 5.5 actor which a structure suitable for cocurrent state management.

import Pioneer

actor Datastore {
    private var users: [ID: User] = [:]

    func find(with ids: [ID]) async -> [User] {
        ids.compactMap { users[$0] }
    }

    func select() async -> [User] {
        users.values.filter(predicate)
    }

    func insert(_ newUser: User) async -> User {
        users[newUser._id] = newUser
        return newUser
    }

    func update(for id: ID, with newUser: User) async -> User? {
        guard let _ = users[id] else {
            return nil
        }

        users[id] = newUser
        return newUser
    }

    func delete(for id: ID) async -> User? {
        guard let user = users[id] else {
            return nil
        }
        users.removeValue(forKey: id)
        return user
    }

    static let shared: Self = .init()
}

This actor should look fairly straightforward.

#Entities

#User profiles

ID type

#User inputs

#Datastore

Contents