Scrapbook

Scrapbook helps you share the things you're working on every day! As a Hack Clubber, you are always learning and building things. Scrapbook allows you to share updates on the things you're doing with the rest of the Hack Club community, and keeps you motivated by recording each day you contribute, tallying that up onto a streak shown on your profile.

Development

Install Bun 1, then:

git clone https://github.com/hackclub/scrapbook.git
cd scrapbook
bun install --frozen-lockfile
bun run generate
bun run dev

Put the web environment in apps/web/.env and the bot environment in apps/slack-bot/.env. Open http://localhost:3000 for the website.

bun run dev:slack-bot     # Bot with Bun's watcher
bun run build            # Generate Prisma client and build the website
bun test                 # Offline bot runtime smoke tests
bun run start            # Serve the built website
bun run start:slack-bot   # Run the bot
bun run --cwd apps/slack-bot checkFormat
bun run --cwd apps/slack-bot format

The repository is a Bun workspace with one root bun.lock. Run installs at the repository root and commit the lockfile when updating dependencies. Scripts explicitly use Bun for Next.js and Prisma so their Node shebangs do not select a separate Node runtime.

Code checks

Oxlint and Oxfmt are installed at the workspace root. ESLint and Prettier are replaced by these scripts:

bun run lint          # Check JavaScript/TypeScript correctness and unused code
bun run lint:fix      # Apply safe lint fixes
bun run format        # Format source, manifests, configs, and workflow files
bun run format:check  # Check formatting without writing
bun run check         # Lint, format check, and offline tests

Oxfmt retains the web app's single quotes / no semicolons and the bot's double quotes / semicolons. The formatter covers source, JSON, YAML, and workflow files. Documentation, CSS, Prisma schemas/migrations, the bot transcript, and generated build files are excluded in .oxfmtrc.json.

Oxlint enables correctness rules plus React and Next.js checks. Native image markup remains allowed for external/user content, and React Compiler purity and synchronous-effect restrictions are not enabled for this existing Pages Router app. Exceptions for trusted transcript templates, the public widget API, manual migration helpers, and the feed's initial cursor effect are documented inline.

GitHub Actions runs independent Lint, Format, Tests, Build (web), and Build (slack-bot) jobs on pushes and pull requests. The build jobs use the actual deployment Dockerfiles and a synthetic BuildKit APP_URL; they do not push images or deploy. Coolify continues to own deployments through compose.yml.

Coolify / Docker Compose

Use the repository's compose.yml with the repository root as the build context. Both Dockerfiles build with dhi.io/bun:1-dev and run with the nonroot, shell-free dhi.io/bun:1 image. If DHI pulls require credentials, configure dhi.io registry credentials on the Coolify build server.

Supply the existing scrappy.env on the deployment server and set PG_DATABASE_URL, GRAPHITE_HOST, and APP_URL in Coolify as before. APP_URL is also needed at build time because Next.js embeds the avatar rewrites. In Coolify, enable its build-time scope and use Docker BuildKit secrets. The Dockerfile reads secret ID APP_URL; the Compose build argument remains a fallback for local builds. Use the hostname without a scheme, matching the existing configuration. Database credentials are runtime-only. Existing port mappings remain web 3002:3000 and bot 3001:3001.

docker compose build

Prisma generation runs during image builds; database migrations remain an explicit operation against the intended database. The image build does not apply migrations. See dependency upgrade notes for the release review and compatibility changes.