#input #action #bevy #mapping #game-engine #gamedev

action_maps

A dynamic action mapping system for Bevy

5 unstable releases

0.3.1 Dec 26, 2023
0.3.0 Dec 24, 2023
0.2.1 Dec 24, 2023
0.2.0 Dec 24, 2023
0.1.0 Nov 23, 2023

#1381 in Game dev

Download history 6/week @ 2024-09-21 1/week @ 2024-09-28

54 downloads per month

MIT license

52KB
1K SLoC

Action Maps

A dynamic action mapping system for Bevy.

Why Action Maps?

Action maps provides an interface similar to Unity's InputSystem and Godot's Input Map. With Action Maps, you can assign functionality to any available Bevy input item without having to work directly with the inputs themselves.

At the moment, this only applies to button type inputs.

Defining Actions

Under the hood, actions only use a String to keep track of their identity. Defining actions is as easy as inserting the name of the action and the input for the action into the control scheme.

fn bind_keys(
    mut control_scheme: ResMut<ControlScheme>
) {
    control_scheme.insert("Jump", KeyCode::Space);
    control_scheme.insert("A", KeyCode::A);
    // or...
    control_scheme.set(make_controls!(
        ("Jump", KeyCode::Space),
        ("A", KeyCode::A),
    ));
}

Handling input

Action maps are wrappers around the existing Bevy Input<T>, so the interface for using them is exactly the same.

use action_maps::prelude::*;

fn handle_input(
    actions: Res<ActionInput>
) {
    if action.just_pressed("Jump") {
        println!("Your character jumps!");
    }
}

A Different Way

Action maps methods all accept any item which implements Into<Action>.

#[derive(Debug, Clone, Copy, PartialEq, Eq, Hash)]
enum Actions {
    Up,
    Left,
    Down,
    Right
}

impl Into<Action> for Actions {
    fn into(self) -> Action {
        match self {
            Actions::Up => Action::from("Up"),
            Actions::Left => Action::from("Left"),
            Actions::Down => Action::from("Down"),
            Actions::Right => Action::from("Right"),
            Actions::ChangeColor => Action::from("ChangeColor"),
        }
    }
}

fn bind_keys(
    mut control_scheme: ResMut<ControlScheme>
) {
    control_scheme.insert(Actions::Up, KeyCode::W);
    // or...
    control_scheme.set(make_controls!(
        (Actions::Up, KeyCode::W),
    ));
}

Universal Input

Action maps provides another layer to help you define keybindings: UniversalInput. UniversalInput is a simple wrapper over every button-type input device available in Bevy. Action maps methods can take any of these input types and use them just the same.

fn bind_keys(
    mut control_scheme: ResMut<ControlScheme>
) {
    control_scheme.insert("Up", KeyCode::W);
    control_scheme.insert("Down", ScanCode(get_scan_code("W")));
    control_scheme.insert("Shoot", MouseButton::Left);
}

Using ScanCodes

Action maps provides a helper function action_maps::get_scan_code to assist in using ScanCodes in Bevy. This function is extremely rudimentary and likely prone to missing keys (not to mention completely lacking support for Linux). Hopefully, it will be made unnecessary with this PR and Bevy 0.13.

let qwerty_w_scancode = action_maps::get_scan_code("W");

Multiplayer

Multiplayer is made easy with the helper types MultiInput and MultiScheme. You can easily import everything you need from action_maps::multiplayer Add in the MultiActionMapPlugin instead to get started.

fn bind_keys(
    mut inputs: ResMut<MultiInput>,
    mut control_schemes: ResMut<MultiScheme>,
) {
    make_multi_input!(
        inputs,
        control_schemes,
        (
            (Actions::Up, ScanCode(get_scan_code("W"))),
            (Actions::Left, ScanCode(get_scan_code("A"))),
            (Actions::Down, ScanCode(get_scan_code("S"))),
            (Actions::Right, ScanCode(get_scan_code("D"))),
        ),
        (
            (Actions::Up, ScanCode(get_scan_code("Up"))),
            (Actions::Left, ScanCode(get_scan_code("Left"))),
            (Actions::Down, ScanCode(get_scan_code("Down"))),
            (Actions::Right, ScanCode(get_scan_code("Right"))),
        )
    );
}


fn handle_input(
    multi_input: Res<MultiInput>,
    mut query: Query<(&mut Transform, &PlayerId)>,
) {
    for (mut transform, PlayerId(id)) in query.iter_mut() {
        let actions: &ActionInput = multi_input.get(*id).unwrap();
        // handle controls here the same way you would for singleplayer!
    }
}

Planned Changes

  • Support for Axis type inputs

Potential Issues

  • Action Maps uses the event system to read input. As such, if you don't ensure the ordering of systems properly, you may end up with a one frame delay between the input being registered and your systems being able to handle the input updated.

Dependencies

~22MB
~433K SLoC