import DiscordBM
import Vapor

let app: Application = <#Your Vapor Application#>
let bot = await BotGatewayManager(
    eventLoopGroup: app.eventLoopGroup,
    httpClient: app.http.client.shared,
    token: <#Your Bot Token#>,
    presence: .init( /// Set up bot's initial presence
        /// Will show up as "Playing Fortnite"
        activities: [.init(name: "Fortnite", type: .game)],
        status: .online,
        afk: false
    ),
    /// Add all the intents you want
    /// You can also use `Gateway.Intent.unprivileged` or `Gateway.Intent.allCases`
    intents: [.guildMessages, .messageContent]
)

In Discord developer portal:

DiscordBM comes with full support for all kinds of “interactions” such as slash commands, modals, autocomplete etc… and gives you full control over how you want to use them using type-safe APIs.

You can see Penny as an example of using all kinds of commands in production.
Penny registers the commands here and responds to them here.

In this example you’ll only make 2 simple slash commands, so you can get started:

In your EntryPoint.main():

/// Make a list of `Payloads.ApplicationCommandCreate`s that you want to register
/// `DiscordCommand` is an enum that has the full info of your commands. See below
let commands = DiscordCommand.allCases.map { command in
    return Payloads.ApplicationCommandCreate(
        name: command.rawValue,
        description: command.description,
        options: command.options
    )
}

/// You only need to do this once on startup. This updates all your commands to the new ones.
try await bot.client
    .bulkSetApplicationCommands(payload: commands)
    .guardSuccess() /// Throw an error if not successful

/// Use the events-stream later since the for-loop blocks the function
let eventsStream = await bot.makeEventsStream()
for await event in eventsStream {
    EventHandler(event: event, client: bot.client).handle()
}

In your EventHandler:

/// Use `onInteractionCreate(_:)` for handling interactions.
struct EventHandler: GatewayEventHandler {
    let event: Gateway.Event
    let client: any DiscordClient
    let logger = Logger(label: "EventHandler")

    /// Handle Interactions.
    func onInteractionCreate(_ interaction: Interaction) async {
        do {
            /// You only have 3 second to respond, so it's better to send
            /// the response right away, and edit the response later.
            /// This will show a loading indicator to users.
            try await client.createInteractionResponse(
                id: interaction.id,
                token: interaction.token,
                payload: .deferredChannelMessageWithSource()
            ).guardSuccess()

            /// Delete this if you want. Just here so you notice the loading indicator :)
            try await Task.sleep(for: .seconds(1))

            /// Handle the interaction data
            switch interaction.data {
            case let .applicationCommand(applicationCommand):
                switch DiscordCommand(rawValue: applicationCommand.name) {
                case .echo:
                    if let echo = applicationCommand.option(named: "text")?.value?.asString {
                        /// Edits the interaction response.
                        /// This response is intentionally too fancy just so you see what's possible :)
                        try await client.updateOriginalInteractionResponse(
                            token: interaction.token,
                            payload: Payloads.EditWebhookMessage(
                                content: "Hello, You wanted me to echo something!",
                                embeds: [Embed(
                                    title: "This is an embed",
                                    description: """
                                    You sent this, so I'll echo it to you!

                                    > \(DiscordUtils.escapingSpecialCharacters(echo))
                                    """,
                                    timestamp: Date(),
                                    color: .init(value: .random(in: 0 ..< (1 << 24) )),
                                    footer: .init(text: "Footer!"),
                                    author: .init(name: "Authored by DiscordBM!"),
                                    fields: [
                                        .init(name: "field name!", value: "field value!")
                                    ]
                                )],
                                components: [[.button(.init(
                                    label: "Open DiscordBM!",
                                    url: "https://github.com/DiscordBM/DiscordBM"
                                ))]]
                            )
                        ).guardSuccess()
                    } else {
                        try await client.updateOriginalInteractionResponse(
                            token: interaction.token,
                            payload: Payloads.EditWebhookMessage(
                                content: "Hello, You wanted me to echo something!",
                                embeds: [Embed(
                                    title: "This is an embed",
                                    description: """
                                    You sent this, so I'll echo it to you but there was nothing!
                                    """,
                                    timestamp: Date().addingTimeInterval(90),
                                    color: .green,
                                    footer: .init(text: "Footer!"),
                                    author: .init(name: "Authored by DiscordBM!"),
                                    fields: [
                                        .init(name: "field name!", value: "field value!")
                                    ]
                                )]
                            )
                        ).guardSuccess()
                    }
                case .link:
                    /// `DiscordBM` has some "require" functions for easier unwrapping of
                    /// application commands. These "require" functions will either give you
                    /// what you want, or throw an error.
                    /// See the full list below.
                    let subcommandOption = try (applicationCommand.options?.first).requireValue()
                    let subcommandName = subcommandOption.name
                    let subcommand = try LinkSubCommand(rawValue: subcommandName).requireValue()

                    let id = try (subcommandOption.options?.first).requireValue().requireString()
                    let name = subcommand.rawValue.capitalized

                    try await client.updateOriginalInteractionResponse(
                        token: interaction.token,
                        payload: Payloads.EditWebhookMessage(
                            content: "Hi, did you wanted me to link your accounts?",
                            embeds: [.init(
                                description: "Will link a \(name) account with id '\(id)'",
                                color: .yellow
                            )]
                        )
                    ).guardSuccess()
                case .none: break
                }
            default: break
            }
        } catch {
            /// Log the errors. Using a `Logger` is preferred compared to just `print()`.
            logger.error("Caught an interaction error", metadata: [
                "error": "\(error)",
                "interaction": "\(interaction)"
            ])
        }
    }
}

In a new file like DiscordCommand.swift, to keep things organized:

// MARK: - Define a nice clean enum for your commands
enum DiscordCommand: String, CaseIterable {
    case echo
    case link

    /// The description of the command that Discord users will see.
    var description: String? {
        switch self {
        case .echo:
            return "Echos what you say"
        case .link:
            return "Links your accounts in Penny"
        }
    }

    /// The options of the command that Discord users will have.
    var options: [ApplicationCommand.Option]? {
        switch self {
        case .echo:
            return [ApplicationCommand.Option(
                type: .string,
                name: "text",
                description: "What to echo :)"
            )]
        case .link:
            return LinkSubCommand.allCases.map { subCommand in
                return ApplicationCommand.Option(
                    type: .subCommand,
                    name: subCommand.rawValue,
                    description: subCommand.description,
                    options: subCommand.options
                )
            }
        }
    }
}

// MARK: - You can use enums for subcommands too 
enum LinkSubCommand: String, CaseIterable {
    case discord
    case github

    /// The description of the subcommand that Discord users will see.
    var description: String {
        switch self {
        case .discord:
            return "Link your Discord account"
        case .github:
            return "Link your Github account"
        }
    }

    /// The options of the subcommand that Discord users will have.
    var options: [ApplicationCommand.Option] {
        switch self {
        case .discord: return [ApplicationCommand.Option(
            type: .string,
            name: "id",
            description: "Your Discord account ID",
            required: true
        )]
        case .github: return [ApplicationCommand.Option(
            type: .string,
            name: "id",
            description: "Your Github account ID",
            required: true
        )]
        }
    }
}

Interaction Parsing Helpers

• StringIntDoubleBool has:
  • requireString() throws -> String
  • requireInt() throws -> Int
  • requireDouble() throws -> Double
  • requireBool() throws -> Bool
• Interaction.ApplicationCommand.Option has all the StringIntDoubleBool functions for unwrapping an option’s value .
• Interaction.ApplicationCommand.Option and related types (like [Option]) have:
  • option(named: String) -> Option?
  • requireOption(named: String) throws -> Option
• [Interaction.ActionRow] and [Interaction.ActionRow.Component] have:
  • component(customId: String) -> Interaction.ActionRow.Component?
  • requireComponent(customId: String) throws -> Interaction.ActionRow.Component
• Interaction.ActionRow.Component has:
  • requireButton() throws -> Button
  • requireStringSelect() throws -> StringSelectMenu
  • requireTextInput() throws -> TextInput
  • requireUserSelect() throws -> SelectMenu
  • requireRoleSelect() throws -> SelectMenu
  • requireMentionableSelect() throws -> SelectMenu
  • requireChannelSelect() throws -> ChannelSelectMenu
• Swift’s Optional has a requireValue() throws function overload (only in DiscordBM).

DiscordBM has support for sending files as attachments.

Note
It’s usually better to send a link of your media to Discord, instead of sending the actual file.

/// Raw data of anything like an image
let image: ByteBuffer = <#Your Image Buffer#>

/// Example 1
try await bot.client.createMessage(
    channelId: <#Channel ID#>,
    payload: .init(
        content: "A message with an attachment!",
        files: [.init(data: image, filename: "pic.png")],
        attachments: [.init(index: 0, description: "Picture of something secret :)")]
        ///                 ~~~~~~~^
        /// `0` is the index of the attachment in the `files` array.
    )
)

/// Example 2
try await bot.client.createMessage(
    channelId: <#Channel ID#>,
    payload: .init(
        embeds: [.init(
            title: "An embed with an attachment!",
            image: .init(url: .attachment(name: "penguin.png"))
            ///                          ~~~~~~~^ 
            /// `penguin.png` is the name of the attachment in the `files` array.   
        )],
        files: [.init(data: image, filename: "penguin.png")]
    )
)

Take a look at testMultipartPayload() in /Tests/DiscordClientTests to see how you can send media in a real-world situation.

DiscordBM contains some utility functions for working with Discord’s text-message format.

The mention helpers:

let userMention = DiscordUtils.mention(id: <#User ID of type UserSnowflake#>)
let channelMention = DiscordUtils.mention(id: <#Channel ID of type ChannelSnowflake#>)
let roleMention = DiscordUtils.mention(id: <#Role ID of type RoleSnowflake#>)
let slashCommandMention = DiscordUtils.slashCommand(name: "help", id: <#Command ID#>)

/// Then:

/// Will look like `Hi @UserName!` in Discord
let userMessage = "Hi \(userMention)!"

/// Will look like `Welcome to #ChannelName!` in Discord
let channelMessage = "Welcome to \(channelMention)!"

/// Will look like `Hello @RoleName!` in Discord
let roleMessage = "Hello \(roleMention)!"

/// Will look like `Use this command: /help` in Discord
let slashCommandMeessage = "Use this command: \(slashCommandMention)"

And the emoji helpers:

let emoji = DiscordUtils.customEmoji(name: "Peepo_Happy", id: <#Emoji ID#>)
let animatedEmoji = DiscordUtils.customAnimatedEmoji(name: "Peepo_Money", id: <#Emoji ID#>)

/// Then:

/// Will look like `Are you happy now? EMOJI` where emoji is like https://cdn.discordapp.com/emojis/832181382595870730.webp
let emojiMessage = "Are you happy now? \(emoji)"

/// Will look like `Here comes the moneeeey EMOJI` where emoji is like https://cdn.discordapp.com/emojis/836533376285671424.gif
let animatedEmojiMessage = "Here comes the moneeeey \(emojiMessage)"

Plus the end-user-localized timestamp helpers:

let timestamp = DiscordUtils.timestamp(date: Date())
let anotherTimestamp = DiscordUtils.timestamp(unixTimestamp: 1 << 31, style: .relativeTime)

/// Then:

/// Will look like `Time when message was sent: 20 April 2021 16:20` in Discord
let timestampMessage = "Time when message was sent: \(timestamp)"

/// Will look like `I'm a time traveler. I will be born: in 15 years` in Discord
let anotherTimestampMessage = "I'm a time traveler. I will be born: \(anotherTimestamp)"

And a function to escape special characters from user input.
For example if you type **BOLD TEXT** in Discord, it’ll look like BOLD TEXT, but using this function, it’ll instead look like the original **BOLD TEXT** input.

let escaped = DiscordUtils.escapingSpecialCharacters("**BOLD TEXT**")

/// Will look like `Does this look bold to you?! **BOLD TEXT**` in Discord. Won't actually look bold.
let escapedMessage = "Does this look bold to you?! \(escaped)"

DiscordBM has the ability to cache Gateway events in-memory, and keep the data in sync with Discord:

let cache = await DiscordCache(
    /// The `GatewayManager`/`bot` to cache the events from. 
    gatewayManager: <#GatewayManager You Made In Previous Steps#>,
    /// What intents to cache their related Gateway events. 
    /// This does not affect what events you receive from Discord.
    /// The intents you enter here must have been enabled in your `GatewayManager`.
    /// With `.all`, `DiscordCache` will cache all events.
    intents: [.guilds, .guildMembers],
    /// In big guilds/servers, Discord only sends your own member/presence info.
    /// You need to request the rest of the members, and `DiscordCache` can do that for you.
    /// Must have `guildMembers` and `guildPresences` intents enabled depending on what you want.
    requestAllMembers: .enabled,
    /// What messages to cache.
    messageCachingPolicy: .saveEditHistoryAndDeleted
)

/// Access the cached stuff:
if let aGuild = await cache.guilds[<#Guild ID#>] {
    print("Guild name is:", aGuild.name)
} else {
    print("Guild not found")
}

DiscordBM has some best-effort functions for checking permissions and roles.
FYI, in interactions, the member field already contains the resolved permissions (Interaction.member.permissions).

Warning
You need a DiscordCache with intents containing .guilds & .guildMembers and also requestAllMembers: .enabled.

let cache: DiscordCache = <#DiscordCache You Made In Previous Steps#>

/// Get the guild.
guard let guild = await cache.guilds[<#Guild ID#>] else { return }

/// Check if the user has `.viewChannel` & `.readMessageHistory` permissions in a channel.
let hasPermission = guild.userHasPermissions(
    userId: <#User ID#>,
    channelId: <#Channel ID#>, 
    permissions: [.viewChannel, .readMessageHistory]
)

/// Check if a user has the `.banMembers` guild-wide permission.
let hasGuildPermission = guild.userHasGuildPermission(
    userId: <#User ID#>,
    permission: .banMembers
)

/// Check if a user has a role.
let hasRole = guild.userHasRole(
    userId: <#User ID#>,
    roleId: <#Role ID#>
)

DiscordLogger enables you to send your logs to Discord with beautiful formatting and a lot of customization options. Read more about it at https://github.com/DiscordBM/DiscordLogger.

React-To-Role helps you assign roles to members when they react to a message. Read more about it at https://github.com/DiscordBM/DiscordReactToRole.

DiscordBM comes with tools to make testing your app easier.

• You can type-erase your BotGatewayManagers using the GatewayManager protocol so you can override your gateway manager with a mocked implementation in tests.
• You can also do the same for DefaultDiscordClient and type-erase it using the DiscordClient protocol so you can provide a mocked implementation when testing.

These are some general implementation detail notes about the DefaultDiscordClient.
Generally, the DefaultDiscordClient will try to be as smart as possible with minimal compromise.

I’ll refer to DefaultDiscordClient as “DDC” to be more concise.

Rate Limits

DiscordBM comes with a HTTPRateLimiter type that keeps track of the x-ratelimit headers.
This, in conjunction with ClientConfiguration‘s RetryPolicy, helps DiscordBM to recover from what that can otherwise be a 429 Too Many Requests error from Discord.
The behavior specified below is enabled by default.

• Before each request, DDC will ask the rate-limiter if the headers allow a request.
• The rate-limiter will respond with yes, you can, no, you can't or yes, but you must wait x seconds first, otherwise no.
• If the response yes, the DDC will continue performing the request.
• If the response no, the DDC will throw a “rate-limited” error.
• If the response yes, but you must wait x seconds first, otherwise no, then the DDC will look at the retryPolicy of its configuration.
• The DDC will act like there has been a 429 error, and will ask the retryPolicy if it’s possible to retry such a failure, and under what circumstances.
• The retryPolicy may specify that 429 requests can be retried basedOnHeaders if not longer than maxAllowed seconds.
• The DDC will wait as long as x seconds which the rate-limiter specified, then will perform the request. This only happens if the x seconds is not longer than the maxAllowed.
• In any other cases other than specified above, the DDC will fail with a “rate-limited” error.

Concurrent Requests

ClientConfigurations CachingBehavior has the ability to avoid multiple concurrent requests with the same “cacheable identity”.

• You can enable caching using DDC’s configuration.cachingBehavior through the initializers by passing cachingBehavior: .minimal, .enabled or the .custom static functions.
• As an example, if you have caching enabled for a cacheable endpoint, and you make 10 concurrent requests to the endpoint with the same parameters, the DDC will only perform 1 of those requests, and let the other 9 requests use the cached value.