@neo9/n9-node-micro

Node MicroService Module

Stats

StarsIssuesVersionUpdatedCreatedSize
@neo9/n9-node-micro
1124.6.12 years ago4 years agoMinified + gzip package size for @neo9/n9-node-micro in KB

Readme

n9-node-micro

Node microservice module: init every module, create an express server, add the required routes (/, /ping & /routes) & module routes.

npm version Travis Coverage license

Installation

npm install --save @neo9/n9-node-micro

Usage

n9Micro([options]): Promise<{ app, server }>

Options:

  • hasProxy: Boolean, default: true, define if the server should handle JWT (see Authentication section below)
  • path: String, default: './modules/'
  • log: Function, default: global.log, need to be a N9Log instance. If no one is found, will use n9Log('n9-node-micro')
  • http: Object, default: {}
    • logLevel: String | false, default: 'dev', see available levels
    • port: Number, default: 5000
    • preventListen: Boolean, default: false
  • jwt: Object, default: { secret: 'secret', expiredIn: '7d' }, only when hasProxy: false
    • secret: String, default 'secret'
    • expiredIn: String, default '7d'

Example

modules/
  users/
    users.init.ts
    users.routes.ts
server.ts

server.ts

import n9Log from '@neo9/n9-node-log'
import n9Micro from '@neo9/n9-node-micro'

(async () => {
  await n9Micro({ log: n9Log('my-app') })
})()

Modules init

Every module can have a *.init.(ts|js) file wich exports a method. This method can be asynchronous (using async or returning a Promise), it reveices an object as argument with log, app and server.

Example:

modules/io/io.init.ts

export default

import * as socketIO from 'socket.io'

export default async function({ log, server }) {
  log.info('Creating socket server')
  global.io = socketIO(server)
}

Routes

Routes format (*.routes.ts) should export an Array of Object with at least these properties:

  • method: String or [String]
  • path: String
  • handler: Function or [Function]

Optional properties:

  • version: String or [String], default: '*'
  • name: String, default: handler.name, for n9-node-connector
  • validate: Object, default: {}
  • session: Boolean or Object, default: false, see Authentication below for more informations.
  • can: Object, check imperium can middleware for usage
  • is: Array<string> | string, check imperium is middleware
  • withAcl: Boolean default: false, ask for acl actions to be injected to the response object. Check imperium for more information
  • documentation: Object, default: {}
    • description: String
    • response: Object, example of response sent back

Example :

modules/foo/foo.routes.js

import * as Joi from 'joi'

export default [
  {
    version: 'v1',
    method: 'post',
    path: '/foo',
    validate: {
      body: Joi.object().keys({
        foo: Joi.boolean()
      })
    },
    documentation: {
      description: 'Foo route',
      response: { ok: true }
    },
    handler: (req, res) => {
      res.json({ ok: true })
    }
  }
]

server.ts

import n9Micro from '@neo9/n9-node-micro'

(async () => {
  await n9Micro()
})()

If we go to http://localhost:5000/routes, we will have:

[
  {
    module: 'foo',
    name: 'handler',
    description: 'Foo route',
    version: 'v1',
    method: 'post',
    path: '/v1/foo',
    session: false,
    acl: [ ],
    validate: {
      body: {
        type: 'object',
        properties: {
          foo: {
            type: 'boolean'
          }
        },
        additionalProperties: false,
        patterns: [ ]
      }
    },
    response: {
      ok: true
    }
  }
]

Application routes

GET - /

Send back the application name (from package.json).

GET - /ping

Send back status code 200 with text pong.

GET - /routes

Send back the list of routes available for the microservice.

Authentication

This module supports 2 ways for authentication, behind a proxy (n9-node-micro-proxy) or alone, to choose which one you want to use, specify the hasProxy option for n9Micro().

  • session header: hasProxy: true
  • JSON Web Token: hasProxy: false

By default, hasProxy is set to true, this means that n9Micro will load req.session from the session header.

When defining the routes in your *.routes.ts files, you can use the session key to define the behaviour of n9Micro.

  • session: Boolean or Object, default: false
    • type: String, default: 'require', values: 'require' or 'load'
    • getToken: Function, default: undefined, called with req as argument and should returns the token to decode (only for hasProxy: false)

When type is 'load', n9Micro will try to load req.session but will not fail.

Session Header

By default, hasProxy is true, this means that n9Micro will load req.session from JSON.parse(req.headers.session).

req.headers.session should be a JSON string and has a userId property when parsed.

JSON Web Token

Only with hasProxy: false

n9Micro supports JWT and will check the Authorization header to be a valid token.

Think to give jwt option to n9Micro() to give the secret key to sign & decode

Example:

// modules/users/users.routes.ts
export default [
  {
    method: 'post',
    path: '/me',
    session: true,
    handler: function(req, res, next) {
      res.json(req.session)
    }
  }
]

Here session: true, will require req.headers.authorization to exists (will use jwt option given to n9Micro() to decode).

The module also add 2 useful methods for loading a session & generating a JWT into req:

  • req.generateJWT(session: Object): Promise<string>

Generates and returns a JWT based on the session.

  • req.loadSession(getToken?: Function): Promise<void>

Will load the session from req and add it to req.session.

getToken is optional, takes req as argument and should returns the JWT (example: (req) => req.query.token)

Example:

// Generate a new token for each authenticated request
app.use(async function (req, res, next) {
  try {
    await req.loadSession() // Add req.session
    res.setHeader('new-token', await req.generateJWT(req.session))
  } catch (err) {
    // Ignore err
  }
})

If you find any bugs or have a feature request, please open an issue on github!

The npm package download data comes from npm's download counts api and package details come from npms.io.