CORUJA technical test

How to install

Create a virtual environment and install dependencies from requirements.txt

virtualenv venv
source venv/bin/activate
pip install -r requirements.txt

Run migrations to create the database and start development server

python manage.py migrate
python manage.py runserver

You may edit coruja_test/settings.py and set GOOGLEMAPS_API_KEY or pass as an environment variable when importing clients to geocode their address.

Now it's ready to run!

Geocoding addresses

In order to get addresses correctly geocoded when importing clients from CSV, you'll need to get a Google Maps api key and provide it as either an environment variable or set it manually in settings.py
If you don't set GOOGLEMAPS_API_KEY, the contacts will be imported but geocoding won't happen, and lat/lon attributes will be set as null. You'll also get a warning.

Import contacts from csv

Set GOOGLEMAPS_API_KEY environment variable and run importcsv management command to import a CSV file. This command takes as its only argument the path of the CSV file to read. (The key below is an example, it wont work.)

GOOGLEMAPS_API_KEY=YOUR_API_KEY python manage.py importcsv customers.csv

Important: Field names must be according to the provided CSV sample customers.csv.

One-command setup

By running the included setup.sh shell script, an interactive bash script does all the setup work.

./setup.sh

Optional: Docker container

Run docker and set the GOOGLEMAPS_API_KEY environment variable. You may not set it, then geocoding won't happen.

sudo docker build -t coruja-test .
sudo docker run -p 8000:8000 -e GOOGLEMAPS_API_KEY=your_key coruja-test 

Please note: The customers.csv file will be imported on the entrypoint. This should take 3-4 minutes. If you want to speed things up, you may replace customers.csv file with a smaller list.

REST API and endpoints

Run development server: manage.py runserver

• Access http://localhost:8000/swagger for the swagger interface

Direct URL For API endpoints are:

• Access http://localhost:8000/api/customer to return a listing of customers
• Access http://localhost:8000/api/customer/[customer_id] to retrieve a specific customer

Misc. comments

• As requested, Google Maps api was used. However, for speed (and cost) optimization, I would suggest another provider such as MapQuest that allows for multiple addresses per query.
• In this demo project, the google maps api key isn't needed to access the REST API endpoints. It is used only by the management command.
• At first, docker container was set to use gunicorn instead of django's development server. But in order to avoid the need of a docker-compose file to serve static content for the swagger interface, I went back to running django's own server for the purpose of this test.
• In production, maybe we shouldn't trust the ID to be unique.