Command Line Application for Finding Nearest Store

Prerequisites

Google API

1. Go to https://developers.google.com/maps/documentation/geocoding/start to create account
2. Select "Geocoding API"
3. Retrieve the API key

Installation

1. Clone repo

$ git clone git@github.com:cinxmo/find-store.git
$ cd find-store/

2. Set environment variables

At the root directory, create an .env file and set

GOOGLE_API_KEY=<Secret API Key>
DEBUG=true

Note: setting DEBUG to false means that the package is not editable (changes to .py modules in find_store after the initial install will NOT be reflected)

3. Install app and dependencies

$ pipenv shell
$ pipenv install

4. Run Application and Usage

$ find_store --address="<address>"
$ find_store --address="<address>" [--units=(mi|km)] [--output=text|json]
$ find_store --zip=<zip>
$ find_store --zip=<zip> [--units=(mi|km)] [--output=text|json]

5. Results

The closest store's name, address, and distance are printed to the console. A file is automatically saved as closest_store.txt unless output is specified as json.

Run tests

$ pipenv install --dev
$ pytest

Assumptions and Explanation

This is a command line application that takes in an address or zip code and prints the name, address, and distance of the closest store listed in store-locations.csv. A text/json file is saved in the root directory depending on the --output option.

Assumptions

While there is a validity check for zip code based on regular expressions, there is none based on address. The assumption is that address input is a string and formatted according to <Address>, <City>, <State> and could include <Zip> at the end as well.

Another assumption is that the user only inputs km or mi for --units, and json or text for --output. If given more time, we should add additional validation functions in util.py.

Distance between two latitude, longitude points is calculated using the Haversine formula. The assumption is that the radius of the Earth is 3958.8 miles.

If given more time, we should increase the test coverage to include:

• Additional bad zip codes
• Calling the Google Geocoder API with invalid/incomplete addresses (ignored this time since calling API for testing could be costly)
• Add additional fixtures for different addresses/zip codes
• Mock errors thrown from Geocoder and handle the errors appropriately so they don't impact user experience