GQL 2 SQL

gql2sql is a teaching-focused example of how to build a modern GraphQL API on top of SQL without stopping at the database. The app exposes books from SQLite through Prisma, combines that with poetry data from a public API, and keeps the codebase structured the way current TypeScript projects are usually built.

What this repo demonstrates

• Apollo Server 5 mounted on Express 5
• Prisma ORM 7 with SQLite and driver adapters
• strict TypeScript with generated GraphQL resolver types
• Biome for formatting and linting
• GraphQL SDL in .graphql files instead of inline strings
• HTTP-level integration tests with Vitest and Supertest
• agent-friendly repository conventions via AGENTS.md

Requirements

• Node 22 or newer
• npm 10 or newer

Architecture

The request flow is intentionally simple and explicit:

1. Express receives HTTP requests and Apollo handles GraphQL execution at /graphql.
2. SDL files in src/graphql/schema define the public contract.
3. Typed resolvers translate GraphQL operations into service calls.
4. Services coordinate domain rules and delegate persistence to repositories.
5. Prisma talks to SQLite for books, while a dedicated poetry client fetches a second upstream data source.

This separation keeps the repo readable for learners and gives coding agents a clear place to make targeted changes.

Project layout

• src/server.ts: startup, env loading, graceful shutdown
• src/app.ts: Express and Apollo composition
• src/config: environment parsing and logging
• src/graphql: SDL files, generated resolver types, and resolvers
• src/services: domain services and the poetry API client
• src/data: Prisma client factory and repositories
• prisma: schema, migrations, and seed script

Quickstart

1. Install dependencies:

   npm install
   

2. Copy the example environment file:

   cp .env.example .env
   

3. Generate the Prisma client, apply the committed migration, and seed the database:

   npm run db:generate
   npm run db:migrate
   npm run db:seed
   

4. Start the development server:

   npm run dev
   

5. Open http://localhost:3000/graphql.

Apollo Sandbox is enabled in development, so you can browse the schema and execute operations directly in the browser.

Example operations

Query all books:

query GetBooks {
  books {
    id
    title
    author
    createdAt
  }
}

Query one book by id:

query GetBook($id: ID!) {
  book(id: $id) {
    id
    title
    author
  }
}

Create a book with a modern input object:

mutation AddBook($input: AddBookInput!) {
  addBook(input: $input) {
    id
    title
    author
  }
}

Variables:

{
  "input": {
    "title": "Kindred",
    "author": "Octavia E. Butler"
  }
}

Query poems from the external provider:

query GetPoetry {
  poem {
    title
    content
    lineCount
    poet {
      name
    }
  }
  poems {
    title
    poet {
      name
    }
  }
}

Scripts

• npm run dev: watch mode with tsx
• npm run build: generate Prisma client, run GraphQL codegen, and compile TypeScript
• npm run typecheck: strict TypeScript verification
• npm run lint: Biome checks
• npm run format: Biome formatting
• npm run test: Vitest suite
• npm run db:generate: Prisma client generation
• npm run db:migrate: apply the committed migration set
• npm run db:reset: reset the local database
• npm run db:seed: seed the database
• npm run db:studio: open Prisma Studio

QA workflow

Run the same checks expected in CI:

npm run lint
npm run typecheck
npm run test

For a clean local reset:

npm run db:reset
npm run db:seed

Notes for contributors and agents

• Keep schema changes in .graphql files and regenerate types with npm run codegen.
• Keep persistence logic out of resolvers.
• If the Prisma schema changes, rerun npm run db:generate and update migrations.
• To author a brand-new Prisma migration locally, use npx prisma migrate dev --name <change> on Node 22 and commit the generated SQL.
• See AGENTS.md for repository-specific coding guidance.