Ecosyste.ms: Awesome

An open API service indexing awesome lists of open source software.

Awesome Lists | Featured Topics | Projects

https://github.com/tlhunter/take-home-test-uber

Coding Challenge for a Taxi Company
https://github.com/tlhunter/take-home-test-uber

Last synced: 25 days ago
JSON representation

Coding Challenge for a Taxi Company

Awesome Lists containing this project

README 

        
# TAXI Coding Challenge



This was my interview code submission for a Taxi company.



## Requirements



This project requires that PostgreSQL be installed.



## Installation



Before using this project, you'll need to have the Node Modules installed. Run the following:



```

npm install

```



Once that is done, you'll need to configure the project to connect to your Postgres database:



```

export PGSQL_CONN="postgres://USER:PASS@HOST/DB"

```



You'll also want to execute the contents of `schema.sql` to create the necessary DB table.



## Sample Data Server



This emits sample data following the criteria outlined in the "Dispatch Backend" document.

It transmits data over a TCP socket, specified as the only argument to the script.



It's not technically part of the project, but the project would be quite boring without it ;).

If you were to actually examine the coordinates, you'd see a lot of drunk Taxi drivers changing directions erratically and driving in the bay.



The service will start emitting simulation data as soon as the first client connects.



```

./emitter.js 2900

```



## Subscription Service



This is the tool which consumes the data.

It also provides an HTTP server for getting data.



Incoming data is immediately put into a PgSQL table.

Two connections are made, one for writes and one for reads, to maximize performance.



```

./service.js 2900 8000

```



### HTTP Endpoints



The service provides three HTTP endpoints for answering the questions asked in the "Dispatch Backend" document.



#### Count Trips within Geo-Rect



This will count all trips which happened within the specified Geo-Rect for the entire known history.



The two points can be any two opposing corners of a rectangle.



	GET /trip-count?p1_lat={latitude1}&p1_lon={longitude1}&p2_lat={latitude2}&p2_lon={longitude2}



	{

	  "trip_count": 200

	}



#### Count Trip Start/Stops and Sum Fares within Geo-Rect



This will count all trips which have started and stopped within the specified Geo-Rect and a sum of their fares for the entire known history.



The two points can be any two opposing corners of a rectangle.



	GET /fare-sum?p1_lat={latitude1}&p1_lon={longitude1}&p2_lat={latitude2}&p2_lon={longitude2}



	{

	  "trip_count": 200,

	  "fare_sum": 100.00

	}



#### Count Concurrent Trips at Specific Time



This will count all distinct trips which were occuring at the specified time.



The timestamp needs to be formatted as ISO 8601.



	GET /snapshot-count?timestamp=2014-03-15T04:02:08Z



	{

	  "trip_count": 300

	}



## Validating the Application



These three commands can be executed to see how quickly the server replies with data.



The third command will require tweaking the timestamp and providing timezone data.



```

time curl http://localhost:8000/trip-count?p1_lat=37.777058&p1_lng=-122.401729&p2_lat=37.763998&p2_lng=-122.380357

time curl http://localhost:8000/fare-sum?p1_lat=37.777058&p1_lng=-122.401729&p2_lat=37.763998&p2_lng=-122.380357

time curl http://localhost:8000/snapshot-count?timestamp=2014-03-16T18:30:00Z

```



## Benchmark Information



On my Quad 2Ghz/8GB/RAM/SSD Debian laptop, with ~300k rows of event data, the SQL queries usually take between 10ms and 20ms to execute.



## Known Issues



If you kill the service and restart it, the data will get out of sync with the "real state" of the world.

So it's best to kill both the server and the emitter, then start the emitter followed by the server.