Hapijs + PostgreSQL Boilerplate

Yet another Hapi boilerplate for a RESTFul server. Opinionated. Using PostgreSQL as Database.

Using Node Native Promises for async handling, standardjs for linting, and grouping files by resource (for the API).

Leverages the hapi ecosystem to a great extent:

• Lab and Code for testing
• Glue for server configuration and plugin registration
• hapi-auth-cookie for authentication
• Hapi builtin authorization via scope (see route options in Hapi API docs for details)
• Confidence for config management
• Joi for validation
• Boom for http error responses
• Good for logging
• Lout (+ Vision & Inert) for automatic documentation

Bookshelf is used as ORM, together with knex (which is also used to run migrations).

This boilerplate has authentication and authorization setup, and user management api routes, for both demo purposes and to bootstrap an application easily.

Project Structure

Hapijs Plugins are used to organize and modularize code. General purpose plugins go into src/ directory. API plugins go into src/api/{resource} directories.

Plugins are wiredeup using Glue. All plugin loading logic is into src/manifest.js file.

For API resources, the directory contains 3 files: index.js, handlers.js and model.js

Example:

src
|- api
   |- users
      |- index.js
      |- handlers.js
      |- model.js

index.js is basically a router file, were plugin is declared and routes are registered into server.

The router handlers are declared in handlers.js file. This file exports a function that receives as argument the options parameter from plugin registration. This serves both to pass global parameters (such as db connections) to handlers and to do dependency injection in tests. An object with all handlers is returned. The handlers signature is allways (request, reply).

const Handlers = require('./handlers')

const options = {
  db: someDBStub
}

let handlers = Handlers(options)

console.log('handlers', handlers)
// {
//    create: function (request, reply) {...},
//    read: function (request, reply) {...},
//    ...
// }

Handlers interact with the Model, where business logic should be encapsulated, such as validations (other than payload validation, which is done at router level), DB operations and such likes. The model file also exports a function which receives the options object passed to Handlers function, and returns an object with all model operations, such as basic CRUD functionality.

All model function return promises.

The model also returns a blueprint object, used for payload validation in router.

Database

PostgreSQL is used as Database, with Bookshelf as ORM and knex as query builder.

Knex is used to run migrations. See knex docs for details on how to create and run migrations.

Configuration

All configuration is stored in src/config.js, using Confidence. Configuration for migrations is in migrations/config.js.

Running locally

• Make sure PostgreSQL is running
• Clone repo
• Run npm install
• Setup environment variables in a .env file (see dotenv). The following environment varibles must be set:
  • NODE_ENV: should be either test, dev, staging or prod
  • COOKIE_SECRET: the secret used to protect the cookie data used in authentication (required only for prod environment)
• Run npm run

Contributing

Feel free to clone and make the changes you want. If you want to commit them back to this repo, please ensure that tests pass and coverage is > 95% (ideally 100%)

License

MIT License