Home
The onebusaway-gtfs-realtime-exporter module provides a number of methods to assist you in producing and sharing a GTFS-realtime real-time transit data feed. Specific features include:
- Support for running a simple, embedded webserver to quickly share your GTFS-realtime data.
- Support for writing GTFS-realtime data to a file.
- Convenience methods for constructing a GTFS-realtime feed.
The onebusaway-gtfs-realtime-exporter module is provided as a Maven module in the OneBusAway Maven Repository. To include it as a dependency to your Maven project, add the following to the <dependencies/> section of your project's pom.xml file:
<dependency>
<groupId>org.onebusaway</groupId>
<artifactId>onebusaway-gtfs-realtime-exporter</artifactId>
<version>1.1.0</version>
</dependency>
How do you use the library? The first step is to write a class that has a dependency on the GtfsRealtimeMutableProvider interface.
@Inject
public void setProvider(GtfsRealtimeMutableProvider provider) {
this.provider = provider;
}
The provider interface defines a contract for providing GTFS-realtime data through three methods:
public interface GtfsRealtimeMutableProvider {
void setAlerts(FeedMessage message);
void setTripUpdates(FeedMessage message);
void setVehiclePositions(FeedMessage message);
}
The three methods correspond to the three GTFS-realtime feed types. The logic to build your GTFS-realtime feed is up to you to provide and will depend on the nature of your underlying real-time system and data-sources.
Once you've built your gtfs-realtime provider class, it's time to wire it up so that we can easily share the data on the web, or optionally write it to a file to share in a more custom way. We make use of the Google Guice dependency-injection framework to wire everything up.
Set<Module> modules = new HashSet<Module>();
GtfsRealtimeExporterModule.addModuleAndDependencies(modules);
Injector injector = Guice.createInjector(modules);
Once we have our Guice injector, we can now configure our application to share GTFS-realtime data via an embedded web-server, a file, or both.
AlertsServlet servlet = injector.getInstance(AlertsServlet.class);
servlet.setUrl(new URL("http://localhost:8080/alerts");
AlertsFileWriter writer = injector.getInstance(AlertsFileWriter.class);
writer.setPath("/some/path/for/alerts");
You can configure all three feed types in a similar manner:
- Alerts: Use the AlertsServlet and AlertsFileWriter
- Trip Updates: Use the TripUpdatesServlet and TripUpdatesFileWriter
- Vehicle Positions: Use the VehiclePositionsServlet and VehiclePositionsFileWriter
Once you've got everything wired up and configured, you can start the application:
LifecycleService lifecycleService = injector.getInstance(LifecycleService.class);
lifecycleService.start();
The GTFs-realtime exporter will now continuously poll your GTFS-realtime provider every few seconds and share the resulting GTFS-realtime data via an embedded web-server or via a written file, as configured above.
By default, when you share a GTFS-realtime feed type through a URL like http://localhost:8080/alerts, the data returned will be binary-encoded Protocol Buffer data, as is expected for a GTFS-realtime feed. However, if you want to see what that data looks like, you can append ?debug to your URL, as in http://localhost:8080/alerts?debug to get a debug dump of the feed that looks something like:
header {
gtfs_realtime_version: "1.0"
incrementality: FULL_DATASET
timestamp: 1330030345776
}
entity {
id: "154323dc759418a69a9e80da7adbc96f"
...
This can be useful when debugging your feed.
Be sure to check out the GtfsRealtimeLibrary class, which has a number of helper methods for constructing GTFS-realtime feeds.
Be sure to check out all our GTFS-realtime resources.