A asset & mod hosting platform for Beat Saber.

To run the server, you must have a instance of PostgreSQL & a database ready for the server to use. If you do not provide Discord OAuth2 credentials, the server will disable authentication and will not allow sign-ins. The importer also requires a Discord bot token in order to make API requests to get users when running the ModelSaber importer. This bot token will also be used to attempt to notify users when they receive an alert, so it is recommended to provide a bot token even if you do not plan on using the importer.

1. Pull the frontend and backend images.
2. Look at the docker-compose.example.yml file for an example of how to set up the server with Docker Compose.
3. Copy the docker-compose.example.yml file to docker-compose.yml and fill in the environment variables.
4. Run docker-compose up -d to start the server.

Both the frontend and backend use yarn as their package manager. You can start both by running yarn dev in their respective folders.

1. Clone the repository.
2. Navigate to the backend folder.
3. Install the dependencies using yarn.
4. Copy the .default.env file to .env and fill in the environment variables.
5. Navigate to the frontend folder.
6. Install the dependencies using yarn.
7. Copy the .default.env file to .env.

The backend server stores all of its data in a folder called storage. If you are using docker, this folder is located at /app/storage. If you are running the server locally, it is located in the root directory of the backend folder (same folder that build & src are). The folder structure for assets is as follows:

storage/
├── uploads/
│   ├── {id}
│   │   ├── {...files}

The uploads folder contains all of the uploaded files. Each upload is stored in a folder named after the upload's ID. Inside each upload folder are the files associated with that upload. Common files will be:

Assets will have both the asset file and up to 5 images in the same folder. The asset file will be named using the fileSafeName property of the asset, and the images will be named according to the iconNames property of the asset. For example, if asset ID 123 has a fileSafeName of example, a type of sound_ogg and iconNames of ["icon1.png", "icon2.png"], the folder for that asset will look like this:

storage/
├── uploads/
│   ├── 123/
│   │   ├── example.ogg
│   │   ├── icon1.png
│   │   ├── icon2.png

Versions of projects will have a similar structure, but the files will be named according to the name property of the project and the version & platform property of the version, as well as being in a sub folder dedicated to that version. For example, if project ID 456 had version ID 532 with a name of Example Mod and a version of 1.0.0 and platform of universal, the folder for that version will look like this:

storage/
├── uploads/
│   ├── 456/
|   |   ├── {project.iconFileName}
|   |   ├── 532/
│   │   |   ├── Example Mod_universal_v1.0.0.zip
|   |   |   ├── Example Mod_universal_v1.0.0_manifest.json
|   |   |   ├── Example Mod_universal_v1.0.0.dll (if the version has been decompiled already)
|   |   |   ├── decompiled/
|   |   |   |   ├── {...decompiled files}

The frontend uses Inlang/Paraglide for translations. All translation files are located in frontend/src/lib/paraglide/messages/. Translations are built when running yarn install in the frontend folder, but you can also run yarn i18n to rebuild them. :::note The frontend does not know what translations it has available, so if you add a new translation, you must also add it to the availableLocales array in frontend/src/lib/stylizer.ts.

The backend uses a custom translation system for user generated content. Errors and alerts generated do not have support for translations.

The backend uses Vitest for testing. Test files are located in the test folder. You can run the tests by executing yarn test in the backend folder. Docker is required for running the tests as it relies on a PostgreSQL container.