We're performing interactive steps adding small things one at a time!
- node 18
mkdir simple-roadmap # create a folder
cd simple-roadmap
npm init -y # creates the `package.json` file, this folder is a node project now
npm i koa
touch index.mjs # you can use just .js extension, but adopting explicit .mjs or .cjs gives more control over module style
This is the initial content of index.mjs:
import Koa from "koa"
const app = new Koa()
app.use(async ctx => ctx.body = "ONLINE")
app.listen(3000)
console.log("http://localhost:3000")
Then add the start script into package.json
:
//...
"scripts" {
"test": "echo \"Error: no test specified\" && exit 1",
"start": "node index.mjs"
}
//...
And run wih npm start
.
Koa is highly modular and there is a dedicated plugin to proper manage routes on it.
- Install koa-router:
npm i @koa/router
Modify index.mjs:
import Koa from "koa"
import Router from "@koa/router"
const app = new Koa()
const router = new Router()
router.get("/status", async ctx => ctx.body = "ONLINE")
app.use(router.routes()).use(router.allowedMethods())
app.listen(3000)
console.log("http://localhost:3000")
Kill previous npm start
and re-run it to see new http://localhost:3000/status
endpoint
- Install level
- Install bodyparser
npm i level
npm i @koa/bodyparser
Modify your index.mjs again:
import Koa from "koa"
import Router from "@koa/router"
import { bodyParser } from "@koa/bodyparser"
import { Level } from "level"
const app = new Koa()
const router = new Router()
const db = new Level("sample", { valueEncoding: "json" })
router.get("/status", async ctx => ctx.body = "ONLINE")
router.get("/todos", async ctx =>
ctx.body = await db.values({ limit: 100 }).all())
router.post("/todos", async ctx => {
const { message, done } = ctx.request.body
const key = new Date().getTime()
const todo = { key, message, done }
await db.put(key, todo)
ctx.body = todo
})
app.use(bodyParser())
app.use(router.routes()).use(router.allowedMethods())
app.listen(3000)
console.log("http://localhost:3000")
Kill previous console again and re-run. Then open a second console and save your first todo:
curl -X POST http://localhost:3000/todos -H 'Content-Type: application/json' -d '{"message":"hello"}'
Check if it was properly saved visiting http://localhost:3000/todos
- Install nodemon so you don't need to kill and restart every time:
npm i -D nodemon
Then modify the scripts section on package.json
:
//...
"scripts" {
"test": "echo \"Error: no test specified\" && exit 1",
"start": "node index.mjs",
"dev": "nodemon index.mjs"
}
//...
For now on, start the program with npm run dev
After add a few more endpoints to complete the REST service, that script will become too horrible to watch:
import Koa from "koa"
import Router from "@koa/router"
import { bodyParser } from "@koa/bodyparser"
import { Level } from "level"
const app = new Koa()
const router = new Router()
const db = new Level("sample", { valueEncoding: "json" })
router.get("/status", async ctx => ctx.body = "ONLINE")
router.get("/todos", async ctx =>
ctx.body = await db.values({ limit: 100 }).all())
router.get("/todos/:key", async ctx =>
ctx.body = await db.get(ctx.params.key))
router.post("/todos", async ctx => {
const { message, done } = ctx.request.body
const key = new Date().getTime()
const todo = { key, message, done }
await db.put(key, todo)
ctx.body = todo
})
router.put("/todos/:key", async ctx => {
const { message, done } = ctx.request.body
const key = ctx.params.key
const todo = { key, message, done }
await db.put(key, todo)
ctx.body = todo
})
router.del("/todos/:key", async ctx =>
ctx.body = await db.del(ctx.params.key))
app.use(bodyParser())
app.use(router.routes()).use(router.allowedMethods())
app.listen(3000)
console.log("http://localhost:3000")
Strictly speaking it works, but it's very coupled and troublesome to test except for integration tests.
Little opportunity for modularization.
Let's start by creating a folder structure and some boilerplate:
mkdir -p app/{controller,service,config}
touch app/controller/todoRequests.mjs app/service/todoService.mjs app/config/db.mjs app/main.mjs
We'll dismantle our single file project into this opinionated folder structure so we can put each concern in it's own place.
Use src
whenever you have any compilation step for your code -- typescript for example.
Use app
folder if it is meant to run the way it is.
import { Level } from "level"
export const db = new Level("sample", { valueEncoding: "json" })
import { db } from "../config/db.mjs"
export const listTodoService = async () =>
await db.values({ limit: 100 }).all()
export const findTodoService = async key =>
await db.get(key)
export const insertTodoService = async ({ message, done }) =>
await updateTodoService({ key: new Date().getTime(), message, done })
export const updateTodoService = async ({ key, message, done }) => {
const todo = { key, message, done }
await db.put(key, todo)
return todo
}
export const delTodoService = async key =>
await db.del(key)
import {
delTodoService,
findTodoService,
insertTodoService,
listTodoService,
updateTodoService
} from "../service/todoService.mjs"
export const listTodoRequest = async ctx => {
ctx.body = await listTodoService()
}
export const findTodoRequest = async ctx => {
const { key } = ctx.params
ctx.body = await findTodoService(key)
}
export const insertTodoRequest = async ctx => {
const { message, done } = ctx.request.body
ctx.body = await insertTodoService({ message, done })
}
export const updateTodoRequest = async ctx => {
const { message, done } = ctx.request.body
const { key } = ctx.params
ctx.body = await updateTodoService({ key, message, done })
}
export const delTodoRequest = async ctx => {
const { key } = ctx.params
ctx.body = await delTodoService(key)
}
import Koa from "koa"
import Router from "@koa/router"
import { bodyParser } from "@koa/bodyparser"
import {
delTodoRequest,
findTodoRequest,
insertTodoRequest,
listTodoRequest,
updateTodoRequest
} from "./controller/todoRequests.mjs"
export const app = new Koa()
const router = new Router()
router.get("/status", async ctx => ctx.body = "ONLINE")
router.get("/todos", listTodoRequest)
router.get("/todos/:key", findTodoRequest)
router.post("/todo", insertTodoRequest)
router.put("/todo/:key", updateTodoRequest)
router.del("/todo/:key", delTodoRequest)
app.use(bodyParser())
app.use(router.routes()).use(router.allowedMethods())
Finally we rewrite the index.mjs once again:
import { app } from "./app/main.mjs"
app.listen(3000)
console.log("http://localhost:3000")
No need to restart the service, nodemon did that for us.
We now have what people call separation of concerns.
npm i -D mocha chai
Create a test spec (app/service/todoService.spec.mjs
):
import * as service from "./todoService.mjs"
import chai, {expect} from "chai"
chai.should()
describe("simple unit test suite", () => {
const message = `message ${new Date().getTime()}`
const messageUpdated = `message ${new Date().getTime()} updated`
let key = -1
it("should create a todo", async () => {
const result = await service.insertTodoService({ message })
result.message.should.be.eql(message)
key = result.key
})
it("should list a todo", async () => {
const result = await service.listTodoService()
result.should.be.an('Array')
})
it("should find a todo", async () => {
const result = await service.findTodoService(key)
result.should.be.an('Object')
result.key.should.be.eql(key)
})
it("should update a todo", async () => {
const result = await service.updateTodoService({ key, messageUpdated })
result.should.be.an('Object')
result.key.should.be.eql(key)
})
it("should delete a todo", async () => {
const result = await service.delTodoService(key)
expect(result).to.be.undefined
})
})
Then modify your test script on package.json
:
//...
"scripts" {
"test": "mocha --recursive app",
"start": "node index.mjs",
"dev": "nodemon index.mjs"
}
//...
Call the tests either with npm run test
or with npx mocha --recursive app
.
Tests are good because having them passing means that the code is supposed to be doing what it should do.
- Install c8
npm i -D c8
Then add a test:coverage script on package.json
:
//...
"scripts" {
"test": "mocha --recursive app",
"test:coverage": "c8 npm run test",
"start": "node index.mjs",
"dev": "nodemon index.mjs"
}
//...
And run it:
npm run test:coverage
This is the sample output:
> simple-roadmap@1.0.0 test:coverage
> c8 npm run test
> simple-roadmap@1.0.0 test
> mocha --recursive app
simple unit test suite
✔ should create a todo
✔ should list a todo
✔ should find a todo
✔ should update a todo
✔ should delete a todo
5 passing (8ms)
-----------------------|---------|----------|---------|---------|-------------------------------
File | % Stmts | % Branch | % Funcs | % Lines | Uncovered Line #s
-----------------------|---------|----------|---------|---------|-------------------------------
All files | 87.39 | 100 | 50 | 87.39 |
app | 100 | 100 | 100 | 100 |
main.mjs | 100 | 100 | 100 | 100 |
app/config | 100 | 100 | 100 | 100 |
db.mjs | 100 | 100 | 100 | 100 |
app/controller | 53.12 | 100 | 0 | 53.12 |
todoRequests.mjs | 53.12 | 100 | 0 | 53.12 | 10-11,14-16,19-21,24-27,30-32
app/service | 100 | 100 | 100 | 100 |
todoService.mjs | 100 | 100 | 100 | 100 |
todoService.spec.mjs | 100 | 100 | 100 | 100 |
-----------------------|---------|----------|---------|---------|-------------------------------
Process finished with exit code 0
Having tests is good, but it's coverage to explain how much we can trust the tests and the code.
- Install chai-http
- Install chai-sinon
- Install sinon
npm i -D chai-http sinon-chai sinon
Create a spec file (app/controller/todoRequests.spec.mjs
):
import chai, {expect} from "chai"
import chaiHttp from "chai-http"
import sinonChai from "sinon-chai"
import * as sinon from "sinon";
import * as controller from "./todoRequests.mjs"
import {app} from "../main.mjs"
import {db} from "../config/db.mjs"
chai.should()
chai.use(chaiHttp)
chai.use(sinonChai)
describe("simple requests test suite", () => {
const sandbox = sinon.createSandbox();
beforeEach(function () {
sandbox.spy(db);
});
afterEach(function () {
sandbox.restore();
});
it("should return 'ONLINE' status", done => {
chai
.request(app.callback())
.get("/status")
.end((err, res) => {
res.text.should.be.eql('ONLINE')
done()
})
})
it("should list todos", (done) => {
chai
.request(app.callback())
.get("/todos")
.end((err, res) => {
res.body.should.be.an("Array")
done()
})
})
it("should insert a todo", async () => {
const ctx = {request: {body: {message: "hello"}}, body: ""}
await controller.insertTodoRequest(ctx)
db.put.should.have.been.calledOnce // sinon-chai in action
})
})
Here we can see chai-http doing some integration tests, and also we can see sinon spying on db calls.
In order to make application more configurable and flexible, we can add checks on environment variables, so we tweak the app behavior accordingly.
We can make listening port configurable:
// index.mjs
import { app } from "./app/main.mjs"
const PORT = process.env.PORT || 3000
app.listen(PORT)
console.log(`http://localhost:${PORT}`)
If PORT environment variable is set, it will be used as listening port.
If no value is set for PORT environment variable, it fallbacks to 3000.
We can make database configurable:
import { Level } from "level"
const LEVELDB = process.env.LEVELDB || "sample"
export const db = new Level(LEVELDB, { valueEncoding: "json" })
console.log(`database is ${LEVELDB}`)
Once the app understands and expects some environment variables it's up to you to properly configure them. Depending on how many projects are present in the developer machine or any other external issue, it might be more tricky than it should be.
On can make use of dot env files to proper manage such variables at development time.
- Install dotenv-flow
npm i dotenv-flow
Then create a file called .env
and add your environment variables:
# variables needed by the application
PORT=3000
LEVELDB=sample
EXTRA_CONFIG=xpto
Finally, you must make the application aware of those variables. To do so, you
need to call the config()
function at entry point, but it's invasive; instead, modify start and dev
scripts in package.json
to perform dynamic loading:
//...
"scripts" {
"test": "mocha --recursive app",
"test:coverage": "c8 npm run test",
"start": "node -r dotenv-flow/config index.mjs",
"dev": "nodemon -r dotenv-flow/config index.mjs"
}
//...
Check if it is working with this change in index.mjs
:
import { app } from "./app/main.mjs"
const PORT = process.env.PORT || 3000
app.listen(PORT)
console.log(`http://localhost:${PORT}`)
console.log(`EXTRA_CONFIG is ${process.env.EXTRA_CONFIG}`)
Kill nodemon process because dynamic loading occurs at startup.
the output should be something like this:
/usr/bin/npm run dev
> simple-roadmap@1.0.0 dev
> nodemon -r dotenv-flow/config index.mjs
[nodemon] 3.0.1
[nodemon] to restart at any time, enter `rs`
[nodemon] watching path(s): *.*
[nodemon] watching extensions: js,mjs,cjs,json
[nodemon] starting `node -r dotenv-flow/config index.mjs`
database is sample
http://localhost:3000
EXTRA_CONFIG is xpto
Let's use gitlab, for GitHub check these examples.
Add the .gitlab-ci.yml file:
image: node:latest
cache:
paths:
- node_modules/
run-tests:
script:
- npm ci
- npm run test:coverage
Now whenever a push is made, gitlab runs the tests.
- Visit https://console.cloud.google.com/run and create a service
- Authorize GitHub (no gitlab, too bad!)
- Select the repository
- Fill the resto of the options, it's pretty straightforward
- Updates on every git push
- Has a free tier
- Visit https://cloud.digitalocean.com/apps
- Authorize Github or Gitlab
- Select the repository
- Updates on every git push
- Costs 12 USD the cheapest plan backend capable
- Visit https://dashboard.heroku.com
- Authorize GitHub
- Select the repository
- Configure Automatic Deploys
- Updates on every git push
- 5 USD the cheapest plan https://www.heroku.com/pricing
- Visit https://dashboard.render.com/
- Authorize GitHub or GitLab
- Select the repository
- Updates on every git push
- Free limited plan and 7 USD ones if you want more than the minimum
under construction