jankboard/README.md
2024-03-03 23:17:38 -08:00

5 KiB

Jankboard 2

maintainability status

A remastered version of the original Jankboard, with a focus on minimizing the Jank portion. Designed from the ground up to be easy to maintain and extend, unlike the original. The frontend is powered by Svelte 4 (unfortunately, we had to drop support for the bleeding edge Svelte 5 due to lack of Vite support). This project is closed source (for now) and licensed under a modified BSD 3.0 license.

Installation guide

We recommend installing the latest stable release binaries from our release page, if you're just looking to run the Jankboard. Alternatively, if you can't find release binaries for your system, follow the directions below in the For developers section to build it from source.

For developers

If you would like to contribute to Jankboard 2, there's only a few simple steps to get the development build up and running.

Prerequisites

  • Rust and cargo. Check the Rust docs for more information. We recommend you install Rust using rustup. Keep in mind that this is ONLY necessary for development, release binaries do not require Rust.
  • NodeJS and npm, for installing dependencies and the vite development server.
  • If you would like the install the (deprecated) Python backend, you will need to install the poetry package manager.
  1. Clone the repository. The app folder contains most of the (soon to be deprecated) Python code that powers the telemetry. The client folder contains the code for the desktop app. The client/src-tauri contains the Rust source code for the backend that interfaces with the robot via Network Tables. The python backend in app is currently being deprecated in favor of a Rust backend. If you need to run it for any reason, check the section below (Troubleshooting common issues).
  2. cd into the client folder. The configuration files and package.json, as well as project code for the desktop app are all stored here. To install dependencies, simply run npm i.
  3. To run the desktop app in developer mode (with automatic hot stateful reload and other useful features like error reporting), make sure you're in the client directory and run npm run tauri dev. This will install and build the Rust dependencies via cargo and initialize the vite development server for the frontend. Note that since this is basically just running the vite development server and then connecting the Tauri webview to it, there may be slightly inconsistent behavior in dev mode versus production mode.
  4. To create a production binary, run npm run tauri build. Tauri cross-compilation is still in beta, so you should generally try to build targeting the same OS you're currently running. Check the Tauri docs for more information.

Troubleshooting common issues

  • If you're experiencing issues with TypeScript type checking or dependency resolution, try opening your editor directly in the client directory so it picks up the tsconfig.json and uses the project's own TypeScript language server.
  • If you don't have access to a development environment that supports running standalone executables (eg. Github Codespaces), you can try running npm run dev instead of npm run tauri dev, which will open a development server at localhost:5173 with the frontend running in the web. However, this may break at any time as critical functionality is more directly attached to the Rust backend.
  • If for some reason you need to install and use the Python backend while we are migrating to Rust, run poetry install --no-root in the root directory of the project to install dependencies. You can start the server with poetry run flask --app app/server.py run --host localhost --port 1280 (it must be running at port 1280 for the frontend to detect it).

Current progress and improvements over (original) Jankboard

  • Layout, toasts/notifications, music player, and app system ported.
  • Toast and audio cue system is much more robust
  • Transitions added almost everywhere to make things smoother
  • Tauri app created successfully, currently still using Flask backend
  • Visualization vastly improved with Threlte (Three.js) powered 3D robot simulation
  • Robot model ported successfully via massive optimization through polygon decimation
  • Added settings app with options to disable certain features and developer tools for testing

TODO

  • Camera cutout overlay
  • Overhaul audio player system
  • Overhaul visualization (especially camera)
  • Overhaul backend in Rust
  • Further integrate telemetry (like GPWS, collision warning, etc)
  • Finish re-creating / adding various voice alerts and sequences
  • Create dynamic voice prompt system to support new languages very easily
  • Add dynamic voice prompt fallback to support incremental voice prompt migration