Mobility Carbon Footprint API

GraphQL API

The main interaction with Mocaf is done over a GrahpQL API. The main API endpoint (and interactive GraphiQL browser) is located at /v1/graphql/.

Authentication

Most GraphQL queries are allowed to access data for only one device. A device is identified by a UUID. Authentication is provided using a randomly generated access token.

The UUID and the access token are passed using a global @device directive for the query (or mutation):

query @device(uuid: "0bb26707-1016-4f60-92e3-62c209040cc2", token: "23730161-54a5-4c06-b0e9-c0533bcc911e") {
  trips {
    id
    legs {
      mode {
        identifier
        name
      }
      length
      carbonFootprint
      startTime
      startLoc
      endTime
      endLoc
    }
  }
}

Mutations

Enable Mocaf

The first interaction with the API should usually be the enableMocaf mutation. It will create the access token and pass it as a response to the mutation.

⚠️ Remember to save the access token! If you lose the access token, you are locked out of the UUID, as all future queries for that UUID will require passing the token.

This mutation is not authenticated, and will fail if a token has already been generated for a device.

Mutation:

mutation {
  enableMocaf(uuid:"4922d7d7-956f-40d0-803f-7d06a0c75d40") {
    ok
    token
  }
}

Response:

{
  "data": {
    "enableMocaf": {
      "ok": true,
      "token": "189c482f-7607-4d12-a0e4-329d9e7e06a9"
    }
  }
}

Disable Mocaf

Calling the disableMocaf mutation will disable the carbon footprint calculator and prevent any other device-specific requests from working except the enableMocaf mutation.

Update leg

The updateLeg mutation allows the user to change leg-specific information such as the transport mode or the number of passengers in the car.

mutation @device(uuid: "0bb26707-1016-4f60-92e3-62c209040cc2", token: "23730161-54a5-4c06-b0e9-c0533bcc911e") {
  updateLeg(leg: "2", mode: "car", nrPassengers: 2) {
    ok
  }
}

Installation

Development

Install the required Python packages:

pip install -r requirements.txt

Create a file called local_settings.py in your repo root with the following contents:

DATABASES = {
    'default': {
        'ENGINE': 'django.contrib.gis.db.backends.postgis',
        'NAME': 'mocaf',
        'ATOMIC_REQUESTS': True,
    }
}

Create a Postgres database and the postgis extension:

createdb mocaf
sudo -u postgres psql -c "CREATE EXTENSION postgis" mocaf

To create the necessary gtfs tables, do the following in the gtfs/sql-importer directory:

PGDATABASE=mocaf make init

If you intend to run the tests, you need to make sure the extension is also created whenever the test database is created, which happens every time you run pytest. If you always run the tests with the --reuse-db flag, you can create the extension once for the database that is being re-used and you'll be fine:

createdb test_mocaf
sudo -u postgres psql -c "CREATE EXTENSION postgis" test_mocaf

If you don't use the --reuse-db flag, the extension needs to be created along with the database every time you run pytest, so you can run the following. Bear in mind, however, that this enables the postgis extension for any new database.

sudo -u postgres psql -d template1 -c "CREATE EXTENSION IF NOT EXISTS postgis"

Also make sure to create the gtfs tables as above.

Run migrations:

python manage.py migrate

Create a superuser:

You might need the following translations during the createsuperuser operation: käyttäjätunnus = username, sähköpostiosoite = e-mail

python manage.py createsuperuser

To import area data for GIS manipulations, use the Django manage.py command import_areas. Please note the area importing has NPM dependencies: before running these commands, run yarn.

Docker Compose

For new deployments with Docker Compose, it should be sufficient to set the PostgreSQL and Cubeserver passwords to something random, and place those variables a .env file:

POSTGRES_PASSWORD=abcdef
CUBEJS_DB_PASS=qwerty

TODO

• Fix viterbi
  • Remove heuristic smoothing
• Incorporate gis for vehicle detection
• Incorporate transitrt_vehiclelocation.route_type in transit type detection
  • E.g. by returning the vehicle_ref
• noProfit