A journey planner which takes into consideration public transport systems, bike sharing services and car sharing / car pooling services to find the best route between a starting and ending point.
This project was developed as a requirement for the course Software Engineering and Development (C02047) at University Of Applied Sciences of Souther Switzerland (SUPSI).
- Giambattista Ravano
- Giancarlo Corti
- Alessandro Trivilini
These are the requirements set by our teacher, therefore most of our choices were made to meet the following requirements
- Insert and connect data of carpooling services such as BlaBlaCar, Carticipate, Cariva
- View car sharing options in a certain area / timespan
- Insert or connect data from public transportation services for a certain area
- Insert data for a transport from one point to another, in a certain area
- Compute the travel proposals, including public transportation systems and carpooling.
- ... other requirements to be decided during the lab activities.
- StarUML
- IntelliJ
- Java and Java FX
Since Java FX is a requirement, we were denied the request of making a web-based application - therefore many of our choices (like the client side routing) were made because of the requirements colliding w/ eachother.
- Java 8+
- JavaFX
- Gradle
Since the project needs to get some public transportation data and route them locally (because it's part of the requirements) and TransitLand doesn't have updated feeds for the Switzerland area (see transitland/transitland#296 for more info about the issue), we had to create and host our own GTFS Server (written in Rust) on gtfs.ded1.denv.it.
The API wrapper for this service is available here.
In order to use a remote GTFS server (default option) the parameter USE_GTFS_REMOTE has to be set to true in src/main/ch/supsi/dti/i2b/shrug/optitravel/config/BuildConfig.java.
The endpoint is hardcoded in the wrapper and is set by default to 127.0.0.1:8000 for local, and gtfs.ded1.denv.it:443 for remote, with HTTPS enabled for remote and disabled for local.
- Docker
- Rust (Nightly)
git clone https://github.com/denysvitali/gtfs-server
cd gtfs-server
Running it in development mode (without the --release flag) won't add some useful optimizations, the import process for the feeds will be much slower!
cargo build --release
docker run --name gtfs-server-db --net gtfs-server-net --ip 172.18.0.2 -e POSTGRES_PASSWORD=mysecretpassword -d mdillon/postgis
./target/release/gtfs-server
If the server starts succesfully, you'll see something similar to the following:
Once the server is started, you can visit http://127.0.0.1:8000/api, if everything works you'll be greeted with a similar page:
You can now head to http://127.0.0.1:8000/import and import the following feeds:
- Switzerland (SBB) feed: https://opentransportdata.swiss/en/dataset/timetable-2018-gtfs/permalink
- London, UK (TfL) feed: https://denv.it/tfl.zip
If you want more feeds, you can get them from http://transitfeeds.com/.
The import process may take a while, and the PostGIS database may grew to more than 4GB, depending on how many feeds you add, and what's the dataset size.
In our tests, importing the SBB and TfL feeds resulted in the following:
basically, a DB of around 4 GB, with 25M of stops, 2M of calendar_dates and 1M of trips. You may now understand why the import process may take this long.
This is why querying for routes in a certain Boudingbox, in a certain timeframe is an intense operation:
we're querying and intersecting 11M + 1M + 5k rows. That's a lot of data.
When the server is up and running, the API is served through the endpoints mentioned in the server documentation, specifically in the routes/api section of the docs.
The project can be started w/
gradle run
If you want to avoid querying the remote services and just have a mocked view, you can
change that in src/main/ch/supsi/dti/i2b/shrug/optitravel/ui/MainController.java by
setting mockedPlanner to true and choosing an appropriate pre-calculated trip by setting the
File f = new File(getClass().getClassLoader()
.getResource("classdata/path-9.classdata").getFile());
to the appropriate classdata file.
Some prebuilt binaries are available on Jenkins. For example, the development builds are available here.
Download the optitravel-*-all.jar and run it with:
java -jar optitravel-*-all.jar
These are the project presentations - part of the course grade.
- Denys Vitali (GTFS Integration, Routing, Planner)
- Alessandro Ferrari (UI)
- Filippo Pura (TransitLand + Routing)
- Zul Arizal (Logo Design)