This is a simple API for finding routes between airports. It uses a modified Dijkstra algorithm to find the shortest routes with additional conditions/restrictions (see below).
The test task consists of two parts, the main part, and a bonus part. We suggest tackling the bonus part once the main objective of the service has been achieved.
The task is to build a JSON over HTTP API endpoint that takes as input two IATA/ICAO airport codes and provides as output a route between these two airports so that:
- The route consists of at most 4 legs/flights (that is, 3 stops/layovers, if going from A->B, a valid route could be A->1->2->3->B, or for example A->1->B etc.) and;
- The route is the shortest such route as measured in kilometers of geographical distance.
For the bonus part, extend your service so that it also allows changing airports during stops that are within 100km of each other. For example, if going from A->B, a valid route could be A->1->2=>3->4->B, where “2=>3” is a change of airports done via ground. These switches are not considered as part of the legs/layover/hop count, but their distance should be reflected in the final distance calculated for the route.
The weekdays and flight times are not important for the purposes of the test task - you are free to assume that all flights can depart at any required time.
- Finding routes using both the flights and switches between adjacent airports.
- Setting a maximum number of stops when finding a route (default is 3).
- Setting a maximum number of 'ground' switches when finding a route (default is 1).
- Some of the API design best practices implemented: logging, CORS, securing HTTP headers, compression, simple validation middleware, and more. For a minimal API with less dependencies, see the
minimalbranch.
To install the project, run npm install.
The API uses airport and route data from the OpenFlights.org public database. The files can also be directly downloaded from GitHub:
The raw data files need to be placed into the data directory with the default filenames. Two sample files are already included in this repository.
Create a file named .env in the project's root directory using the .env.example file as example/template. Modify values for the environment variables as desired, or keep the default values.
Run npm run seed to process the raw data and output (seed) data files used by the API.
To run the project, run npm start. This will start the server on the port defined in the .env file (3000 by default).
The API exposes an endpoint at /flights/find with the following query parameters:
from: (Required) IATA or ICAO code for the source airportto: (Required) IATA or ICAO code for the destination airportmax_stops: (Optional, default value: 3) Maximum number of stops/layovers, not including stops for 'ground' switches (see Task description)max_switches: (Optional, default value: 1) Maximum number of 'ground' switches between adjacent airports
http://localhost:3000/flights/find?from=TLL&to=PSP
http://localhost:3000/flights/find?from=TLL&to=PSP&max_stops=1
http://localhost:3000/flights/find?from=TLL&to=PSP&max_stops=1&max_switches=2
http://localhost:3000/flights/find?from=TLL&to=KDL&max_switches=2
http://localhost:3000/flights/find?from=EETN&to=YSGE&max_stops=5
A successful response has the following format:
{
result: {
airports: Information about all airports across the resulting route,
from: Source airport ID,
to: Destination airport ID,
segments: Resulting route segments,
totalDistance: Total distance in kilometers
}
}
Sample response:
{
"result": {
"airports": {
"415": {
"id": 415,
"name": "Lennart Meri Tallinn Airport",
"city": "Tallinn-ulemiste International",
"country": "Estonia",
"iata": "TLL",
"icao": "EETN",
"latitude": 59.41329956049999,
"longitude": 24.832799911499997
},
"737": {
"id": 737,
"name": "Stockholm-Arlanda Airport",
"city": "Stockholm",
"country": "Sweden",
"iata": "ARN",
"icao": "ESSA",
"latitude": 59.651901245117,
"longitude": 17.918600082397
},
"3484": {
"id": 3484,
"name": "Los Angeles International Airport",
"city": "Los Angeles",
"country": "United States",
"iata": "LAX",
"icao": "KLAX",
"latitude": 33.94250107,
"longitude": -118.4079971
},
"3839": {
"id": 3839,
"name": "Palm Springs International Airport",
"city": "Palm Springs",
"country": "United States",
"iata": "PSP",
"icao": "KPSP",
"latitude": 33.8297004699707,
"longitude": -116.50700378417969
}
},
"from": 415,
"to": 3839,
"segments": [
{
"from": 415,
"to": 737,
"distance": 390.55,
"type": "FLIGHT"
},
{
"from": 737,
"to": 3484,
"distance": 8863.04,
"type": "FLIGHT"
},
{
"from": 3484,
"to": 3839,
"distance": 175.92,
"type": "FLIGHT"
}
],
"totalDistance": 9429.51
}
}A route segment's type can be one of the following:
FLIGHTGROUND(switches between adjacent airports)EITHER(the segment can be covered either with a flight or a ground switch)
An error response has the following format:
{
error: Error information
}
Sample response:
{
"error": "Cannot find airport with the code 'TALL'."
}To run tests, run npm test.
Tests use a subset of the data provided in the data directory: airports and routes in the Baltic states. Four flight routes have been added for testing purposes: two routes between Kärdla (KDL) <-> Kuressaare (URE) and two between Kuressaare (URE) <-> Ruhnu (EERU). That is, these routes will later be marked with the EITHER type. See the testing graph (flight: red, ground switch: gray, either: red and white):
The implementation of this API is significantly based on the following projects: