Make The Pick

A real-time web app for drafting and trading Pokemon with friends. Create a league, draft your team live, negotiate trades, and compete in challenge runs — all scored by a configurable rules engine.

Built as a proof-of-concept for a larger "draft & trade anything" platform.

Prerequisites

• Deno (v2+)
• Docker (for PostgreSQL)

Getting Started

1. Start everything:

   deno task dev

   This single command handles the full setup:

   • Creates .env from .env.example (if not present)
   • Installs dependencies (deno install)
   • Starts PostgreSQL via Docker Compose
   • Launches the server (Deno + tRPC) and client (React + Vite) in parallel

2. Stop the database when you're done:

   deno task db:stop

Available Tasks

Task	Description
deno task dev	Start both server and client in dev mode
deno task test	Run all tests (server + client)
deno task setup	Copy .env.example to .env and install deps
deno task build	Build the client for production
deno task start	Start the server in production mode
deno task db:start	Start PostgreSQL via Docker Compose
deno task db:stop	Stop PostgreSQL
deno task db:generate	Generate Drizzle ORM migrations
deno task db:migrate	Run database migrations
deno task sync:pokemon	Sync Pokemon data from PokeAPI
deno task test:e2e	Run Playwright end-to-end tests

E2E Tests

End-to-end tests use Playwright and run the full app in production mode against a separate database.

Running tests

Make sure PostgreSQL is running (deno task db:start), then:

deno task test:e2e

That's it. On first run, Playwright's global setup automatically creates the make_the_pick_e2e database and runs migrations. Subsequent runs reuse the database and only apply new migrations.

You'll need Playwright browsers installed once:

cd e2e && npx playwright install chromium

How it works

• The app boots in production mode — Vite builds the client, Hono serves it as static files alongside the API on port 3000.
• Auth is handled by seeding user and session rows directly into the E2E database and injecting the session cookie into the browser context. No server-side auth bypass — this is the same mechanism Google OAuth produces at the end of a real login.
• The E2E database is truncated between test suites to keep tests isolated.

In CI

The e2e job in .github/workflows/ci.yml runs after unit and integration tests pass. It spins up its own Postgres service, installs Chromium, runs migrations, and executes the tests. On failure, the Playwright HTML report is uploaded as a build artifact.

Environment Variables

Configuration lives in .env (created automatically from .env.example on first run). Key variables:

Variable	Description
DATABASE_URL	PostgreSQL connection string
BETTER_AUTH_SECRET	Secret for Better Auth sessions
BETTER_AUTH_URL	App URL for auth callbacks
GOOGLE_CLIENT_ID	Google OAuth client ID
GOOGLE_CLIENT_SECRET	Google OAuth client secret
BETTER_AUTH_TRUSTED_ORIGINS	Comma-separated list of trusted origins

Status

Early development — see the technical proposal for the full vision and architecture.

License

MIT