Given the size of the assignment, having everything in the same file was
OK. However, splitting into different files is a good option for better
maintainance once this API grows, and now is a good time to do it before
it gets more complicated.
The database that was chosen here is SQLite, because it's dead simple to
setup and more than enough for this project.
Please note that I took some liberty with the assignment. I chose to use
a numeric field for the `id` column of an item. This leverages automatic
creation and incrementation of the id by SQLite itself.
This initial version is purposely extremily limited:
* `GET /ahoy` endpoint, just to check the server is alive
* `GET /items` endpoint, always returns the same item
* `POST items` endpoint, checks and print request payload, no
persistance
* calling any other route defaults to a 404
