Weather Stats (for Turfcoach)

Minimalistic Python script to prompt a shell user for cities and print weather stats for them.

Requirements

An API key from openweathermap.org

1. Sign up for an OpenWeather account
2. Confirm your email address
3. Go to API keys and copy the default key (or create a custom one)

Installation

1. Clone this repository

   git clone {paste-url}

2. Add your API key as an environment variable

   export OPENWEATHER_API_KEY={YOUR API KEY}

3. Install dependencies

   poetry install

Usage

1. Run the script

   poetry run python turfcoach/weather.py --city "San Francisco" --city Berlin --city Dublin --city Lima

Design Considerations

1. Minimal filestructure: One could argue to extract the code in weather.py into all kinds of files. With the scope as it is, I believe the simplicity of one file outweighs the benefits of a more nuanced file structure.

2. Poetry: I went for Poetry over pyenv, venv, virtualwrapper, etc simply because I hadn't tried it yet and was curious.

3. Async: For a larger project of this kind I would strongly consider async or some coroutine library, not only to parallelize fetch requests for muliple cities. As it is, this seemed the simplest way to the target.

4. Env variables: Usually I'd put sensitive data into an .env file (and exclude that from VCS) or use the secret management of the cloud provider.

5. Other bells and whistles: I considered adding custom exceptions, tests, a license, .gitignore, and a proper frontend but, again, decided for the MVP version.

Additional libraries

• click for a minimal command-line interface
• requests for convenient API calls (instead of urllib.request)
• plotext for printing a histogram to the command line

While a large-scale application might benefit from pandas, I deliberately went for dictionaries for the sake of simplicity.