Todo list using a Sinatra REST API

I am attending to a post-degree program and one of its courses is Server-Side Scripting. The professor gave us a project which students should create a project and my colleague and I created a Sinatra REST API for a todo list application.

The back end is a decent example of how to use Sinatra and Active Record to create simple APIs. In the repository’s readme, I show how to install it and use it. In this post, I detail what I coded:

Managing dependencies and configuring the database

First, I need to specify the dependencies for this project. Here are the gems:

ruby '2.2.2'

source 'https://rubygems.org'

gem 'sinatra'
gem 'sinatra-cross_origin'
gem 'json'
gem 'activerecord'
gem 'pg'
gem 'sinatra-activerecord'
gem 'rake'

group :development do
  gem 'shotgun'
end

In this project, I use shotgun to update my Sinatra app without restarting the server every time that I change my app.rb. The gem sinatra-cross_origin is needed to allow me perform requests externally. In addition, I use activerecord as ORM and sinatra-activerecord to extends Sinatra with extension methods and Rake tasks.

The file environment.rb specifies the database credentials:

configure :production, :development do
  set :show_exceptions, true

  db = URI.parse(ENV['DATABASE_URL'] || 'postgres://127.0.0.1/todo')

  ActiveRecord::Base.establish_connection(
    adapter:  db.scheme == 'postgres' ? 'postgresql' : db.scheme,
    host:     db.host,
    username: db.user,
    password: db.password,
    database: db.path[1..-1],
    encoding: 'utf8'
  )

  ActiveRecord::Base.class_eval do
    def self.reset_autoincrement(options={})
      options[:to] ||= 1
      case self.connection.adapter_name
        when 'MySQL'
          self.connection.execute "ALTER TABLE #{self.table_name} AUTO_INCREMENT=#{options[:to]}"
        when 'PostgreSQL'
          self.connection.execute "ALTER SEQUENCE #{self.table_name}_id_seq RESTART WITH #{options[:to]};"
        when 'SQLite'
          self.connection.execute "UPDATE sqlite_sequence SET seq=#{options[:to]} WHERE name='#{self.table_name}';"
        else
      end
    end
  end
end

I created a reset_autoincrement method that will be used in my seeds file. I will talk about it soon. I also created a cors.rb file that enables Cross Domain Resource Sharing (CORS).

configure do
  set :allow_origin, :any
  set :allow_methods, [:get, :post, :options, :delete, :put]

  enable :cross_origin
end

options "*" do
  response.headers["Allow"] = "HEAD,GET,PUT,POST,OPTIONS,DELETE"
  response.headers["Access-Control-Allow-Headers"] = "X-Requested-With, X-HTTP-Method-Override, Content-Type, Cache-Control, Accept"
end

Last not the least, I need to create a Rakefile that will be used on my rake tasks:

require 'sinatra/activerecord/rake'

namespace :db do
  task :load_config do
    require "./app"
  end
end

Defining models

Diagram

As you can see in ERD above, my todo list has two entities: List and Task.

class Task < ActiveRecord::Base
  belongs_to :list

  validates :name, presence: true
  validates :list_id, presence: true
end

class List < ActiveRecord::Base
  has_many :tasks, dependent: :destroy

  validates :name, presence: true
  validates :color, presence: true, format: /\A#?(?:[A-F0-9]{3}){1,2}\z/i
end

I used the validates method to ensure some valid data. The attribute color, for example, must be a hexadecimal color. The next step is creating the migrations:

$ rake db:create_migration NAME=lists

This command creates a file inside db/migrate. This is my migration file:

class Lists < ActiveRecord::Migration
  def up
    create_table :lists do |t|
      t.string :name
      t.string :color
    end
  end

  def down
    drop_table :lists
  end
end

The same applies to Task:

$ rake db:create_migration NAME=tasks
class Tasks < ActiveRecord::Migration
  def up
    create_table :tasks do |t|
      t.string :name
      t.references :list, foreign_key: true
    end
  end

  def down
    drop_table :tasks
  end
end

After defining the table schema, it’s time to create the tables on the database:

$ rake db:migrate

Creating the API routes

Here is the initial version of my app.rb:

require 'sinatra'
require 'sinatra/cross_origin'
require 'sinatra/activerecord'
require './config/environments'
require './config/cors'
require './models/list'
require './models/task'
require 'json'

before do
  content_type :json
end

get '/lists' do
  List.all.to_json(include: :tasks)
end

get '/lists/:id' do
  List.where(id: params['id']).first.to_json(include: :tasks)
end

post '/lists' do
  list = List.new(params)

  if list.save
    list.to_json(include: :tasks)
  else
    halt 422, list.errors.full_messages.to_json
  end
end

put '/lists/:id' do
  list = List.where(id: params['id']).first

  if list
    list.name = params['name'] if params.has_key?('name')
    list.color = params['color'] if params.has_key?('color')

    if list.save
      list.to_json
    else
      halt 422, list.errors.full_messages.to_json
    end
  end
end

delete '/lists/:id' do
  list = List.where(id: params['id'])

  if list.destroy_all
    {success: "ok"}.to_json
  else
    halt 500
  end
end

The routes for Task entity are very similar:

get '/tasks' do
  Task.all.to_json
end

get '/tasks/:id' do
  Task.where(id: params['id']).first.to_json
end

post '/tasks' do
  task = Task.new(params)

  if task.save
    task.to_json
  else
    halt 422, task.errors.full_messages.to_json
  end
end

put '/tasks/:id' do
  task = Task.where(id: params['id']).first

  if task
    task.name = params['name'] if params.has_key?('name')

    if task.save
      task.to_json
    else
      halt 422, task.errors.full_messages.to_json
    end
  end
end

delete '/tasks/:id' do
  task = Task.where(id: params['id'])

  if task.destroy_all
    {success: "ok"}.to_json
  else
    halt 500
  end
end

I added two extra routes to this application. One will render a index.html file, which I describe the project. The another one will populate the database with my initial data:

get '/' do
  content_type :html
  send_file './public/index.html'
end

get '/refresh' do
  # Clean the database and create the initial data
  load './db/seeds.rb'
end

Time to create the config.ru file, responsable for start the application:

require './app'
run Sinatra::Application

Everything is done, let’s go:

$ shotgun

Hopefully, shotgun should return something like:

== Shotgun/WEBrick on http://127.0.0.1:9393/
[2016-08-03 19:51:50] INFO  WEBrick 1.3.1
[2016-08-03 19:51:50] INFO  ruby 2.2.2 (2015-04-13) [x86_64-darwin14]
[2016-08-03 19:51:50] INFO  WEBrick::HTTPServer#start: pid=13348 port=9393

After that, you will be able to do requests as I did:

Seeding initial data

The file db/seeds.rb populates the database with initial data:

Task.delete_all
List.delete_all
Task.reset_autoincrement
List.reset_autoincrement

puts 'Creating sample lists'
colors = ['54b2a1', '95d5cf', '809bbe', '98d2f3', '80bf86', 'a3d49f']
['Todo', 'Movies', 'Supermarket'].each_with_index do |list, index|
  List.find_or_create_by(name: list, color: colors[index])
end

puts 'Creating sample tasks'
['Nathan\'s Assignment', 'Go to Meetup'].each do |task|
  Task.find_or_create_by(name: task, list: List.where(name: 'Todo').first)
end

['The Godfather', 'Star Wars'].each do |task|
  Task.find_or_create_by(name: task, list: List.where(name: 'Movies').first)
end

['Milk', 'Bread', 'Butter'].each do |task|
  Task.find_or_create_by(name: task, list: List.where(name: 'Supermarket').first)
end

Deployment – Heroku

In order to deploy the app on Heroku, create a Procfile:

web: bundle exec ruby app.rb -p $PORT

Another option is clicking on the “Deploy to Heroku” button on the repository page. Heroku will look for the app.json file and setup all that you need. If you area curious about this file:

{
  "name": "todo-api",
  "description": "A simple Sinatra REST API",
  "keywords": ["sinatra", "api", "activerecord", "reminders"],
  "repository": "https://github.com/leonardofaria/todo-api",
  "addons": [
    "heroku-postgresql:hobby-dev"
  ],
  "env": {
    "RACK_ENV": "production"
  },
  "scripts": {
    "postdeploy": "bundle exec rake db:migrate && bundle exec rake db:seed"
  }
}

This is not the focus of this post, then read the Heroku documentation [1, 2] if you are interested on this. Pretty interesting.

Final thoughts

I am not sure if I described everything but I hope that you got the main idea. You can take a look in the source code on Github, fork it or star it.

A web app (gif above) was also created to show how to use the API. The web app is also available in Github.

Other references

Part of my work was inspired in the following articles:

 

Leonardo

Registros de Leonardo Faria sobre desenvolvimento web