Poo Clouds is implemented entirely in Swift and SwiftUI. Additionally, it follows The Composable Architecture, which allows for clean, concise, truly declarative view code, while making the entire app deeply unit testable. As an example, I have separated out the piece of code that makes the CDC API call as a TCA Dependency. By doing that, I'm able to fully unit test all of the data response handling and model building code that runs when the API call returns simply by providing a mock dependency override for the data fetching function that simply returns an example response. Every line of production code in this scenario that is not a network request is fully tested. Amazing!

The data for the app comes directly from the US National Wastewater Surveillance System and is hosted by the CDC. I made use of their soda-swift package, but it seems to be a pretty thin wrapper around making vanilla REST API calls. I wrapped their callback based network fetching function with a continuation so that my app code could use it with Swift's async/await syntax instead.

Apart from really digging into TCA and seeing first-hand all the benefits, I was also really impressed with the Swift Charts API. The graph in the second screenshot above is a mere 30-ish lines of generously whitespaced, declarative, and extremely readable code.

Unfortunately, the Apple App Store has a (very understandable) policy that apps containing COVID-19 information must come from organizations that have a history of providing accurate medical information - not individual app developers like myself. I am currently shopping around to find a suitable organization (perhaps the CDC itself, which already has a number of apps in the Apple App Store) to which I can donate this code for distribution.