Skip to main content

Creating commands

Commands represent the main form of interaction that Discord bots have with users. Creating commands is a simple and straightforward process, with the ability to customize different aspects of them.

Chat input (aka "slash") commands#

Basic command#

A slash command is a command that is triggered when the user sends /command-name in chat. In the library, they are called "chat input commands".

Here is an example of a /ping command that makes the bot reply with "Pong!":

package testbot1;
import botrino.interaction.annotation.ChatInputCommand;
import botrino.interaction.listener.ChatInputInteractionListener;
import botrino.interaction.context.ChatInputInteractionContext;
import org.reactivestreams.Publisher;
@ChatInputCommand(name = "ping", description = "Pings the bot to check if it is alive.")
public final class PingCommand implements ChatInputInteractionListener {
@Override
public Publisher<?> run(ChatInputInteractionContext ctx) {
return ctx.event().createFollowup("Pong!");
}
}
  • A chat input command must have a @ChatInputCommand annotation that contains the meta-information required by Discord (name of the command, description, defaultPermission, etc), and must implement the ChatInputInteractionListener interface.
  • The run method accepts a ChatInputInteractionContext that holds contextual information on the command being executed, such as the original ChatInputInteractionEvent, the MessageChannel where the interaction happened, the User who initiated the interaction, and a Locale that may have been adapted to the target user (see Filtering and adapting events).
  • Events are automatically acknowledged by default, so you can directly call createFollowup() without using deferReply() first (reply() will not work unless you disable automatic acknowledgment, see Acknowledging Interactions)
info

If you are using the Botrino framework, you have nothing else to do, the command will be automatically recognized and registered. Otherwise, you need to manually register it into the InteractionService like this:

interactionService.registerChatInputCommand(new PingCommand());

Command options#

A command may accept one or many options, whether they are required or optional. The library provides ChatInputCommandGrammar that allows to inject the option values into the fields of a class that is going to be instantiated when the command is executed. Here is an example of a command using options:

package testbot1;
import botrino.interaction.annotation.ChatInputCommand;
import botrino.interaction.context.ChatInputInteractionContext;
import botrino.interaction.grammar.ChatInputCommandGrammar;
import botrino.interaction.listener.ChatInputInteractionListener;
import discord4j.core.object.command.ApplicationCommandOption;
import discord4j.discordjson.json.ApplicationCommandOptionData;
import org.reactivestreams.Publisher;
import java.util.List;
@ChatInputCommand(name = "options", description = "Option testing")
public class OptionsCommand implements ChatInputInteractionListener {
private final ChatInputCommandGrammar<Options> grammar = ChatInputCommandGrammar.of(Options.class);
@Override
public Publisher<?> run(ChatInputInteractionContext ctx) {
return grammar.resolve(ctx.event()).flatMap(options -> ctx.event()
.createFollowup("Value of `my-string`: " + options.myString));
}
@Override
public List<ApplicationCommandOptionData> options() {
return grammar.toOptions();
}
private static class Options {
@ChatInputCommandGrammar.Option(
type = ApplicationCommandOption.Type.STRING,
name = "my-string",
description = "The string argument",
required = true,
choices = {
@ChatInputCommandGrammar.Choice(name = "Choice 1", stringValue = "1"),
@ChatInputCommandGrammar.Choice(name = "Choice 2", stringValue = "2"),
@ChatInputCommandGrammar.Choice(name = "Choice 3", stringValue = "3")
}
)
String myString;
}
}
  1. Create a new class that declares the fields in which you want to inject the option values. It is recommended to use an internal class for better code readability, unless you are re-using the same class for several commands. The class must have a no-arg contructor, and must be declared static if internal. The fields may be private, but you can also omit the access modifier to avoid IDE warnings about the field never being assigned a value.
  2. Use the annotation @ChatInputCommandGrammar.Option on the field to declare the properties of the option (the type, the name, the description, whether they are required or not, and the array of value choices, if any).
  3. Create a new ChatInputCommandGrammar and pass the class to the .of() method. You only need to instantiate once, rather than on each command execution.
  4. In the run(ChatInputInteractionContext) method, call the resolve(ChatInputInteractionEvent) method which will read the options, instantiate the class and inject the values in the annotated fields. You can then use the object to conveniently access the values, as show in the example above.
  5. Override the options() method from ChatInputInteractionListener and make it return ChatInputCommandGrammar#toOptions().

For reference, here is a table associating each ApplicationCommandOption.Type with the type of the field carrying the annotation:

Option typeType of annotated field
STRINGjava.lang.String
INTEGERjava.lang.Long (primitive long may be used only if required = true)
NUMBERjava.lang.Double (primitive double may be used only if required = true)
BOOLEANjava.lang.Boolean (primitive boolean may be used only if required = true)
USERdiscord4j.core.object.entity.User (or discord4j.core.object.entity.Member if in a guild)
CHANNELdiscord4j.core.object.entity.channel.Channel
ROLEdiscord4j.core.object.entity.Role
MENTIONABLEdiscord4j.common.util.Snowflake
caution

Optional options will be filled with null if not specified by the user, which means you cannot use primitive types for INTEGER, NUMBER and BOOLEAN if required = false, otherwise you will get NullPointerExceptions.

Subcommands and subcommand groups#

Discord allows to create subcommands and subcommand groups to help in organizing the logic of a complex command. Here is an example of a command using subcommands and subcommand groups:

@ChatInputCommand(
name = "nest",
description = "Subcommand testing",
subcommands = {
@Subcommand(name = "sub1", description = "Subcommand 1", listener = NestCommand.Sub1.class),
@Subcommand(name = "sub2", description = "Subcommand 2", listener = NestCommand.Sub2.class)
},
subcommandGroups = {
@SubcommandGroup(name = "group1", description = "Group 1", subcommands = {
@Subcommand(name = "sub", description = "Subcommand", listener = NestCommand.G1Sub.class)
})
}
)
public final class NestCommand {
public static final class Sub1 implements ChatInputInteractionListener {
@Override
public Publisher<?> run(ChatInputInteractionContext ctx) {
return ctx.event().createFollowup("sub1: pong!");
}
}
public static final class Sub2 implements ChatInputInteractionListener {
@Override
public Publisher<?> run(ChatInputInteractionContext ctx) {
return ctx.event().createFollowup("sub2: pong!");
}
}
public static final class G1Sub implements ChatInputInteractionListener {
@Override
public Publisher<?> run(ChatInputInteractionContext ctx) {
return ctx.event().createFollowup("group1 sub: pong!");
}
}
}

Here are the notable differences:

  • The class carrying the @ChatInputCommand annotation no longer implements ChatInputInteractionListener. Indeed, as per Discord's documentation a base command becomes unusable if subcommands are present.
  • The @ChatInputCommand specifies an array of @Subcommand and @SubcommandGroup with their own name and description.
  • Subcommands specify the class implementing ChatInputInteractionListener that is going to handle them. In this example they are internal classes, but they can as well be external.
info

Here is how you manually register a command containing subcommands when you control the instance of InteractionService:

interactionService.registerChatInputCommand(new NestCommand(), List.of(
new NestCommand.Sub1(),
new NestCommand.Sub2(),
new NestCommand.G1Sub()
));
caution

If you are using the Botrino framework, the subcommand classes must either have a public no-arg constructor or be declared as a service. If the classes are internal, they must be static.

Context menu commands#

Discord currently support two types of context menu commands, one on messages and one on users. It works the same as chat input commands, but you need to use the @MessageCommand and @UserCommand annotations with the MessageInteractionListener and UserInteractionListener interfaces, respectively.

Context menu commands are actually less complex than chat input ones, since there is no description, no options, no subcommands... Only a name and a run method:

package testbot1;
import botrino.interaction.annotation.UserCommand;
import botrino.interaction.context.UserInteractionContext;
import botrino.interaction.listener.UserInteractionListener;
import org.reactivestreams.Publisher;
@UserCommand("Fight")
public final class FightCommand implements UserInteractionListener {
@Override
public Publisher<?> run(UserInteractionContext ctx) {
return ctx.event().createFollowup("You are fighting <@" +
ctx.event().getTargetId().asString() + ">");
}
}
info

If you need to do manual registration, it happens via InteractionService#registerMessageCommand(MessageInteractionListener) and InteractionService#registerUserCommand(MessageInteractionListener):

interactionService.registerUserCommand(new FightCommand());

Commands as a service#

info

The following is only applicable if you are using the Botrino framework. See Working with services.

Classes implementing commands can themselves be declared as services without any issues. For example if you need to access the ConfigContainer in your command, you can do this:

package testbot1;
import botrino.api.config.ConfigContainer;
import botrino.api.config.object.BotConfig;
import botrino.interaction.annotation.ChatInputCommand;
import botrino.interaction.context.ChatInputInteractionContext;
import botrino.interaction.listener.ChatInputInteractionListener;
import com.github.alex1304.rdi.finder.annotation.RdiFactory;
import com.github.alex1304.rdi.finder.annotation.RdiService;
import discord4j.gateway.intent.IntentSet;
import org.reactivestreams.Publisher;
@RdiService
@ChatInputCommand(name = "intents", description = "Displays the intents enabled for this bot.")
public final class IntentsCommand implements ChatInputInteractionListener {
private final long intents;
@RdiFactory
public IntentsCommand(ConfigContainer configContainer) {
this.intents = configContainer.get(BotConfig.class)
.enabledIntents()
.orElse(IntentSet.nonPrivileged().getRawValue());
}
@Override
public Publisher<?> run(ChatInputInteractionContext ctx) {
return ctx.event().createFollowup("Intents enabled: " + intents);
}
}

The command above accesses the values in the config.json to get the gateway intents enabled for the bot. You can notice the use of @RdiService on top of @ChatInputCommand, this works totally fine! Don't forget the @RdiFactory to inject the configuration container, and you're ready to run the bot and try out this command.

tip

If you declare a command as a service this way, you are allowed to do anything with it like any other service, for example inject it in other services, or set up @RdiFactory to be a reactive static method in case the command needs to perform a reactive task in order to be initialized.