1 unstable release

0.1.0 Jun 24, 2020

#32 in #event-stream

Apache-2.0

140KB
2.5K SLoC

Live data export tool for ActyxOS

ActyxOS is a decentralized event sourcing system for the factory shop-floor. One of the purposes of deploying apps in factories is to gain business insights by reporting and visualizing what is going on on the shop-floor.

With this library it is straight-forward to export event data into SQL databases with code like the following:

pub fn dashboard_logic<'a>(
    scope: &mut Scope<'a>,
) -> (
    Input<Event<MachineEvent>>,
    Flow<'a, DashboardEntry, Stateful>,
) {
    let (injector, events) = Flow::<Event<MachineEvent>, _>::new(scope);

    let out = events
        .filter(|ev| ev.stream.name.as_str().starts_with("Drill"))
        .map(|ev| match ev.payload {
            MachineEvent::Started { order } => {
                DashboardEntry::working(ev.stream.name.to_string(), order, ev.timestamp)
            }
            MachineEvent::Stopped { .. } => {
                DashboardEntry::idle(ev.stream.name.to_string(), ev.timestamp)
            }
        })
        .group_by(|entry| entry.machine.clone())
        .max_by(|entry| entry.since)
        .ungroup();

    (injector, out)
}

The full project is in the example folder, the sample webapp that generates the underlying events can be found in the webapp folder. Both can be deployed to an ActyxOS node or run in developer mode, see the quickstart guide.

See also the blog post for more background information.

Building the Docker app

Build the binary and docker image on Linux with

cd machine-dashboard-app
./build.sh

then package it into an ActyxOS app and deploy it with

ax apps package
ax apps deploy --local com.actyx.data_flow.machine-dashboard-1.0.0-x86_64.tar.gz localhost
ax apps start --local com.actyx.data_flow.machine-dashboard localhost
ax logs tail --local -f localhost

Further steps

From the example you can go and explore more dataflow transformations offered by the library as well as connect to PostgreSQL or Microsoft SQL Server databases to store the transformation results.

The DB driver code is also commented and allows extensions for other databases.

If you want to build on Windows or for other architectures, you may take a look at the Rust Docker image.

Dependencies

~45–61MB
~1M SLoC