AR LavaFlow

Repository structure

The production-ready part of this project is split into two folders: wasm and svelte-app. The wasm folder contains the model generation and smoothing code, which was written in Rust for performance reasons. svelte-app is our frontend, which contains the image processing, and the Augmented Reality visualization of the generated model.

Other than that, in the beginning of the project, we created a lot of smaller demos. These proof-of-concepts can be found in poc.

Development setup

To run this project, the following dependencies need to be installed on your system:

• Rust toolchain with wasm32-unknown-unknown target
• wasm-pack
• NodeJS (v16 LTS or newer recommended) with yarn package manager

Then to run a local development server, run the following commands:

cd svelte-app
wasm-pack -v build ../wasm --target web
yarn
yarn dev

Note that you will need to manually rerun the wasm-pack and yarn dev commands whenever you change the Rust code. If this gets annoying, you could use a tool such as watchexec:

cd svelte-app (if you were not already)
watchexec -cr -w wasm/src/ -- "wasm-pack -v build ../wasm --target web && yarn dev"

Deploying to production

This repository is currently being mirrored to GitHub, where a pipeline is run to deploy it to GitHub pages. But if this changes in the future, you can run follow the instructions for setting up a development server, but run yarn build instead of yarn dev.

This will build a static website in svelte-app/build/, which can be deloyed on a webserver of choice. Though keep in mind that GitHub pages requires some further magic, such as adding a .nojekyll file to make sure folders starting with an underscore are served normally.

---

AR LavaFlow (c) by PRIME, TU Delft

AR LavaFlow is licensed under a Creative Commons Attribution 4.0 International License.

You should have received a copy of the license along with this work. If not, see http://creativecommons.org/licenses/by/4.0/.