Vapi is an API that gathers Norway's digital roadnet, then segments it into reasonably sized road segments. We also accept input for production data related to road condition influences like weather and maintenance (snow plowing, gravelling, salting). This data is mapped to the road segments, and outputted as a RESTful API.
Vapi was built by a team of seven students in their sixth semester of the Bachelor’s in Informatics programme at NTNU, as a part of the course IT2901 - Informatics Project II, colloquially known as the bachelor's thesis. Vapi was developed for the Norwegian Public Roads Administration.
Additional documentation and information can be found in our wiki. As this is a student project, the wiki also contains a lot of administrative information related to the project.
Vapi was built with Django as a multi-container Docker application. Huge thanks to Christian Duvholt for Docker guidance.
Thanks to Jan Kristian Jensen for letting us use his repository for interacting with the NVDB API. We used it for getting the official road network from the NVDB API and processing it.
We recommend cloning the repository with SSH (see here for adding SSH key if you haven't already):
[email protected]:IT2901-24-2018/vapi.git
cd vapi
If you wish to contribute to Vapi, we would also appreciate you setting the following configuration options:
git config --global core.autocrlf false
git config --global user.name "<your github username>"
git config --global user.email [email protected]
Docker (Compose) will handle most of our dependencies, but first you need to install it! Install the Developer Desktop for Mac or Windows, or the Linux Server for your distribution of choice here.
The Developer Desktop for Mac and Windows already include Docker Compose, but Linux users will have to install Compose here.
Our Django container depends on environment variables. Make a file called .env at the root directory, with the following contents:
DJANGO_DEBUG=True
We have a Makefile that simplifies interaction with our Docker containers. You will likely need to use sudo for most Docker commands, but you can omit it depending on your installation. On Linux you can remove the need for sudo by adding your current user to the docker group:
sudo gpasswd -a $USER docker
After logging out and in again, you'll be able to run Docker (and make) commands without sudo.
Start up the application for the first time, building your containers and starting them up by running:
make
Your Postgres container might be slower than the Django container when starting Vapi for the first time, but the server isn't ready yet anyways, so that's hardly an issue. Stop the server with CTRL+C.
Get your database up to speed by applying existing migrations:
make migrate
Create a superuser in Django:
make superuser
You are now ready to start the server again:
make
Navigate to http://localhost:8000/, and see that Vapi is up and running!
To gain access to all the endpoints of the API, you must log in as a superuser. Navigate to the API page and log in with your superuser in the top-right corner.
We have two files to help with populating the database with road segments and production data. These scripts are not necessary for the API, but we left them here as a convenience for future developers.
To run our example scripts, you will need to fill in your superuser credentials to the .env file you made earlier. This file will not be checked in to Git. Modify the .env file so that it looks like this, replacing yourusername with your username, and yourpassword with your password:
DJANGO_DEBUG=True
API_USERNAME=yourusername
API_PASSWORD=yourpassword
To run our example road segmenter example_roadnet_to_db.py, make sure that the server is running. Open a new terminal, then make a bash-shell in your container with:
docker exec -t -i vapi_django_1 /bin/bash
When you're inside, run:
python /vapi/apps/data/road_segmenting/example_roadnet_to_db.py
You can also run example_create_test_prod_data.py to simulate production data input here, which requires you to have a apps/data/production_input_data.geojson file:
python /vapi/apps/data/example_create_test_prod_data.py
Our road_fetcher.py file interacts with the NVDB API directly, and is used by both our example_roadnet_to_db.py script, and the test_segmenting.py tests.
Please fill in your client info that is sent along with the requests to the NVDB API, so that they have information on who is using the API.
Make the nvdbapi-clientinfo.json file in the root folder by:
cp example-nvdbapi-clientinfo.json nvdbapi-clientinfo.json
Then, fill in your system and contact into the nvdbapi-clientinfo.json file. The file is ignored by git.
Not having a nvdbapi-clientinfo.json file will produce a warning, but road_fetcher.py will still work.
Default is the first command defined in the Makefile, and as such will be run by simply running make
| Target | Command | Description |
|---|---|---|
| default | build start | Both builds and starts the containers. |
| build | docker-compose build | Builds the containers. |
| start | docker-compose up | Starts the server. |
| down | docker-compose down | Tears down the containers, handy for clearing out the database. |
| stop | docker-compose stop | Stops your containers if they're already running mode. |
| restart | stop start | Stops and starts (restart) your containers. |
| migrate | docker-compose run --rm django python manage.py migrate | Applies existing migrations to the database |
| migrations | docker-compose run --rm django python manage.py makemigrations | Creates new migrations based on changes made to models. |
| superuser | docker-compose run --rm django python manage.py createsuperuser | Initiates Django superuser creation. |
| psql | docker-compose exec -u postgres postgres psql | Starts a psql shell, handy for manually interacting with the database. |
| test | test-only lint-only | Runs our linting and tests. We advice you to run this before pushing your code to Github, to ensure your code will pass our Travis CI build. |
| lint-only | docker-compose run --rm django flake8 apps/ docker-compose run --rm django isort -c |
Only runs linting (flake8 and isort) |
| test-only | docker-compose run --rm django py.test --cov-config=setup.cfg --cov=apps | Only runs our own tests |