feat: add container support

[0xDEC0DE]

Add a Dockerfile and a docker-compose.yaml to allow for easier spin-up and development of a sandboxed environment

Uses uv in lieu of pip for installing Python packages because uvs installer is (much) faster and has a (much) better solver

fix: update dependency versions

Move away from Hatch's "most-compatible version" notation to using plain minimum-version notation with explicit upper-bounds where needed to mitigate breaking changes

[0xDEC0DE]

Pros

• Super simple to set up
• Highly reproducible
• Excellent sandboxing — can wildly experiment without worrying about trashing the local system
• Local vs. "public" MQTT server is dead simple to configure
• Moves "network performance as a bottleneck" to the setup phase only
• The built image can be pushed to Docker Hub, and subsequently be used without needing to build anything

Cons

• Requires Docker/containerd/etc. which can be intimidating
• Excellent sandboxing — changes inside the container will be lost unless explicitly saved, and changes outside the container (may) require a rebuild to be used at all
• If pushing to Docker Hub, Intel and ARM architectures will need to be built and pushed
• The built ARM image is ~1GB compressed, which may be a lot to ask of a fragile public library wifi connection
• The built Intel image is ~5GB compressed, due to the Nvidia CUDA drivers
• Building the image is similarly onerous, needing to fetch gigabytes of data from Debian, PyPI, and GitHub

[0xDEC0DE]

So...

The service attempts to bind to on port 5000 on the host, because that is the default value of the Flask development server and I never bothered to change it, however this will NOT run on a Mac "out of the box", because the Airplay Receiver also binds to that port.

I missed this because, like most people that ever do any amount of Flask dev, I turned off the Airplay Receiver years ago and forgot about it. This should move the exposed port to something unused, e.g., 5888.