SnakeDeck 🐍🎛️

SnakeDeck (or snakedeck, as you like) is a program written in Python to manage devices like the Elgato Stream Deck.

The code quality is barely alpha level; treat it like a prototype or MVP. That being sai, I use it "in production" (to control my lights, cameras, and a few more things necessary when I run my Docker and Kubernetes training courses).

Why?

I wrote SnakeDeck because streamdeck-ui didn't have the features that I needed. SnakeDeck is, however, using the same underlying Python library to interface the Stream Deck.

What can I do with it?

The most interesting feature of SnakeDeck is probably the fact that you can associate Python code to the Stream Deck keys, and that Python code can return the new state of the keys (for instance, to reflect some "on/off" state). It is also (relatively) easy to update key status remotely.

Getting started

Clone that repo, pip install --user -r requirements.txt; and you might have to do some setup specific to the Stream Deck library (FIXME add link to that).

Then run snakedeck.py. If it detects a Stream Deck, it will display a message like the following one:

WARNING:root:Deck AL012345678 has no configuration file (/home/jp/.config/snakedeck/AL012345678.yaml).

We need to create that configuration file. SnakeDeck monitors that location, so we don't need to restart SnakeDeck after creating the file. Furthermore, after loading the file, SnakeDeck keeps monitoring the file, and automatically reloads it when it changes.

(We still need to restart SnakeDeck when changing its code or its plugins, though. For now, at least.)

The configuration file should be a list of key bindings.

Key bindings

A key binding will look like this:

- line: 1
  column: 1
  label: Hello, world!
  emoji: 😎
  shell: curl http://localhost:5000
  eval: print(42)

The line and column is the position of the key on the Stream Deck. (Line 1 is on top, column 1 is on the left.)

You probably want to specify label or emoji, not both. The text of the label will be shrunk if it doesn't fit on the key (so if you put a super long label it will be in tiny text). The emoji will be zoomed to take the whole key. Emojis require to have the Noto Color Emoji font installed.

If the key as a shell section, that command will be executed in a shell.

If the key has an eval section, it will be evaluated as Python code. It can be anything you like (including dangerous stuff that would totally mess up with SnakeDeck's internal state, because why not).

If the code in the eval section returns something (that is not None), that something should be a dict and will be used to update the state of the key. This lets us change the label, the emoji, or even the action associated with the key.

Cycles

A key binding can also be a cycle, like this:

- line: 1
  column: 1
  cycle:
  - emoji: 🙂
  - emoji: 🤣
  - emoji: 😭
  - emoji: 🤬

Each time it is pressed, the key will change to the next state in the cycle, and will loop back to the beginning at the end. In the example above, the keys don't do anything (other than changing what's shown on them) but the keys can of course have actions associated with them, like in the example below.

- line: 1
  column: 1
  cycle:
  - emoji: 🔇
    shell: pactl set-sink-mute @DEFAULT_SINK@ off
  - emoji: 🔊
    shell: pactl set-sink-mute @DEFAULT_SINK@ on

Sync

A key can be synchronized with other Stream Decks, like this:

- line: 1
  column: 1
  cycle:
  - emoji: ✅
  - emoji: ⛔️
  sync: stopgo

When a synchronized key is pressed, SnakeDeck starts sending multicast packets every second with the state of the key. The packet contains the synchronization tag (in that case stopgo).

When SnakeDeck receives one of these packets, and has a key with the same synchornization tag, it updates the state of the key. Note that it completely overwrite the state of the key.

Packets also contain a serial number that gets incremented automatically when the state of the key changes.

The synchronization mechanism is currently bidirectional, but might change in the future to allow send-only and receive-only operation.

Plugin initialization

It is possible to run arbitrary code when the configuration is loaded, by adding an entry with an eval key, without line/column information. For instance, the lights plugin (to interface with Elgato Key Lights) will use Zeroconf to discover existing lights, but in some circumstances, lights might not be detected; and it's possible to manually add a light by adding the following section in the configuration:

- eval: lights.add("10.0.0.42")

Future developments

Feel free to open issues if you have great ideas of new features!