bevy_query_service

bevy_query_service is a bevy library that aims to help with handling requests that requires an input and expects an output (like HTTP calls).

The motivation comes from the limitation of events in Bevy, where events are one-way triggers that leaves the user hanging, whether or not the event is successfully triggered and what's the outcome of the event.

Usage

To use bevy_query_service, simply add it to your Cargo.toml or run cargo add bevy_query_service.

To add a new request, you need a few elements:

Request and reply component

The request and reply are defined as components, to be bundled with the entity.

#[derive(Component, Clone)]
struct Request
{
    /* … */
}

#[derive(Component, Clone)]
struct Reply
{
    /* … */
}

Request event wrapper

An QueryEvent needs to be added to the system for the request to go through.

app.add_event::<QueryEvent<Request>>();

Request server

To generate a reply, a run_query_server needs to be added.

app.add_systems(Update, run_query_server::<Request, Reply>);

To calculate the reply, a QueryReplyOps trait need to be defined.

impl QueryReplyOps<Request> for Reply {
    fn get_reply(world: &mut World, request: &QueryRequest<Request>) -> Self {
        /* … */
    }
}

The get_reply() function has access to the World, which allows easy access to all the entities and resources in the application. This should help users to get all the information they need to formulate a reply.

Methodology

The methodology of bevy_query_service is to use entities to store the requests and process them in the application, before marking them as compeleted.

In each entity, there is a GoalComponent:

#[derive(Component, Debug, Clone, Default)]
pub struct GoalComponent {
    pub uuid: uuid::Uuid,
    pub is_completed: bool,
    pub to_delete: bool,
    pub timer: bevy_time::Stopwatch,
}

The uuid variable allows any system in the application to send a request and retrive the reply using the query function of Bevy.

is_completed helps to keep track if the reply component is constructed while to_delete marks the entity to be deleted by the optional cleanup_requests system. The cleaning up action is kept optional because users might want to implement their own cleanup systems.

The timer component is a work-in-progress. The aim is to have a timer that automatically drops the request after a certain timeout.

Features

• options of feedback