Conceptually, I tried a Clean-architecture approach:

┌───────────────────────────────┐
│ cmd/                          │  → entrypoints (web, seeder)
└───────────────▲───────────────┘
                │
┌───────────────┴───────────────┐
│ internal/server               │  → wiring / composition root
└───────────────▲───────────────┘
                │
┌───────────────┴───────────────┐
│ internal/handler              │  → HTTP adapters, handlers, and renderer
│   ├── middleware              │
│   └── renderer                │
└───────────────▲───────────────┘
                │
┌───────────────┴───────────────┐
│ internal/service              │  → application logic (use cases)
└───────────────▲───────────────┘
                │
┌───────────────┴───────────────┐
│ internal/repo                 │  → infrastructure adapters
│ internal/infra                │  
└───────────────▲───────────────┘
                │
┌───────────────┴───────────────┐
│ internal/domain               │  → core model
└───────────────────────────────┘

Use golang 1.25 Libraries:

• github.com/go-chi/chi/v5
• github.com/gorilla/csrf
• github.com/gorilla/sessions
• github.com/jackc/pgx/v5
• golang.org/x/crypto/argon2
• golang-migrate

Run $ go mod vendor to resolve references

Database can be deployed using podman/docker on development folder. ~/development$ docker compose up -d Database that is used is PostgreSQL. PostgreSQL is used here with these reasons:

• I expect the schema of this project to evolve given time.
• I rely on complex queries such as to aggregate. This means that I pushed logic near the data related to it.
• This is the safer long-term database because its extensibility. Who knows I might need it to use extension like PostGIS (geospatial) or simply use its capability for full-text search and query-able JSONB data type.

Check Makefile

• To create migration: $ make create-migration name=new_migration
• To migrate all pending migration: $ make migrate
• To migrate down one previous migration: $ make migrate-down
• To seed initial users: run $ make seed-database OR run manually $ DB_URL="postgres://here" go run ./cmd/seed/main.go. Seed file can be done using separate .csv files:
  • cmd/seed/users.csv contains seed data for users table
• To generate new mocks $ make mockery
• To test the code $ make test

Environment variables are located in file .env-example. This file eventualy will be .env on the deployemnt environment or loaded through different manners.

ENV="development"
PORT=":3000"
DB_URL="postgres://postgres:password@localhost:5433/backend_db?sslmode=disable"
COOKIE_SECRET="Rm57qySVRliOZg5WqJ5GyKHKY6f4sJ41"
CSRF_SECRET="4eQWYCt7WjxLwPmL06MhOW5FS96wxOk6"
JWT_SECRET="4eQWYCt7WjxLwPmL06MhOW5FS96wxOk6"
UPLOAD_DIR="../../uploads"
HOSTNAME="http://localhost:3000"

• TBD

APIs are designed to utilize JSON:API specification JSON:API Spec. At some point, simple json should be enough.

Return the page to login via web.

Route to handle the authentication logic and return cookies and JWT. Primarily used by the form from the login page. The resulting token can be used to authenticate future endpoints.