• If you are citizen of an European Union member nation, you may not use this service unless you are at least 16 years old.

  • Stop wasting time looking for files and revisions. Connect your Gmail, DriveDropbox, and Slack accounts and in less than 2 minutes, Dokkio will automatically organize all your file attachments. Learn more and claim your free account.


Todos 06-Building with Sinatra and DataMapper

Page history last edited by Eric Retzloff 10 years, 1 month ago

About Sinatra, DataMapper, and Heroku


Sinatra is a very simple Ruby framework for generating web applications.  Thanks to its small size and tight design, it makes a great backend for building the kind of simple REST-oriented services that drive SproutCore applications.


DataMapper is a fast Ruby-based ORM for managing data in a database.  We'll be using DataMapper to interface with a SQLite database for storage in this tutorial.


Before You Get Started


Before you start coding with Sinatra, you will need to make sure it is installed on your machine.  Enter the following command to get all the Ruby modules you need:


sudo gem install sinatra data_mapper data_objects do_sqlite3 dm-sqlite-adapter json


This will install or update all of the packages we'll be using.


Create Your App


NOTE: If you want to skip writing the code for this yourself, just install the ruby gems above then checkout the samples-todos-sinatra Git repository at Github.


Most Sinatra applications reside in a single Ruby file.  This one will be no different.  To setup your app, just create a new ruby file and call it "tasks.rb".  The entire app is included below, with comment:


in tasks.rb:

#!/usr/bin/env ruby


# Implements Tasks spec defined at: 

# http://wiki.sproutcore.com/Todos+06-Building+the+Backend


require 'rubygems'

require 'sinatra'

require 'dm-core'

require 'dm-migrations/adapters/dm-sqlite-adapter'

require 'json'


# connect DataMapper to a local sqlite file. 


DataMapper::setup(:default, ENV['DATABASE_URL'] || 

    "sqlite3://#{File.join(File.dirname(__FILE__), 'tmp', 'tasks.db')}")


# define the Task model object we will use to store data

# in the server.  Note the three properties defined.  "id"

# will be used as the GUID for the task.


class Task 

  include DataMapper::Resource


  property :id,          Serial

  property :description, Text, :required => true

  property :is_done,     Boolean


  # helper method returns the URL for a task based on id  


  def url




  # helper method converts the Task to json.  Anytime you

  # call to_json on a data structure with a Task, this will

  # be used to convert the task itself


  def to_json(*a)


      'guid'        => self.url, 

      'description' => self.description,

      'isDone'      => self.is_done 




  # keys that MUST be found in the json

  REQUIRED = [:description, :is_done]


  # ensure json is safe.  If invalid json is received returns nil

  def self.parse_json(body)

    json = JSON.parse(body)

    ret = { :description => json['description'], :is_done => json['isDone'] }

    return nil if REQUIRED.find { |r| ret[r].nil? }







# instructs DataMapper to setup your database as needed




# return list of all installed tasks.  Just get all tasks and

# return as JSON

get '/tasks' do

  content_type 'application/json'

  { 'content' => Array(Task.all) }.to_json



# create a new task.  request body to contain json

post '/tasks' do

  opts = Task.parse_json(request.body.read) rescue nil

  halt(401, 'Invalid Format') if opts.nil?


  task = Task.new(opts)

  halt(500, 'Could not save task') unless task.save


  response['Location'] = task.url

  response.status = 201



# Get an individual task

get "/tasks/:id" do

  task = Task.get(params[:id]) rescue nil

  halt(404, 'Not Found') if task.nil?


  content_type 'application/json'

  { 'content' => task }.to_json



# Update an individual task

put "/tasks/:id" do

  task = Task.get(params[:id]) rescue nil

  halt(404, 'Not Found') if task.nil?


  opts = Task.parse_json(request.body.read) rescue nil

  halt(401, 'Invalid Format') if opts.nil?


  task.description = opts[:description]

  task.is_done = opts[:is_done]



  response['Content-Type'] = 'application/json'

  { 'content' => task }.to_json



# Delete an invidual task

delete '/tasks/:id' do

  task = Task.get(params[:id]) rescue nil

  task.destroy unless task.nil?





That's all there is to it.  Put the above file into tasks.rb and then start your server with the following command:


$ ruby ./tasks.rb


You should now be able to test that your server works.  We recommend using HTTPClient if you are on a Mac since it makes it easy to construct HTTP requests and see their results.   


Here are some basic HTTP actions you can try along with the expected results:


GET /tasks


This should return a list of tasks in JSON.  If you try this first, the list of tasks should be empty.  Once you've created a few tasks, the response should be something like:


HTTP/1.1 200 OK

Content-Type: application/json

Content-Length: 72

Connection: close

Server: thin 1.2.4 codename Flaming Astroboy


{"content":[{"isDone":true,"guid":"/tasks/8","description":"edit me "}]}


POST /tasks


When you send this HTTP request, make sure you include task JSON in the body.  The post body should look something like this:


"POST /tasks" post body:

{ "description": "test task", "isDone": true }


The response from your post should be empty except for a status code of 201 "Created" and a "Location" header with the new URL for the created task:


HTTP/1.1 201 Created

Location: /tasks/9

Content-Type: text/html

Content-Length: 0

Connection: close

Server: thin 1.2.4 codename Flaming Astroboy


GET /tasks/(id)


Once you've created a task, you can try retrieving it.  Just GET the URL returned by the Location header.  The return value should be a content hash just like in /tasks:


HTTP/1.1 200 OK

Content-Type: application/json

Content-Length: 69

Connection: close

Server: thin 1.2.4 codename Flaming Astroboy


{"content":{"isDone":true,"guid":"/tasks/9","description":"test task"}}


PUT /tasks/(id)


This command should update the task.  Like POST /tasks, you need to pass a post body with your new task.  The post body should contain both the description and isDone properties to be valid.  Be sure to PUT to a URL representing an existing task (like the one you created previously):


"PUT /tasks/(id)" post body:

{ "isDone": false, "description": "EDITED" }


The return value should indicate 200 OK and should also return the new task JSON.  You should now be able to go GET /tasks/(id) and see the properties you just modified change.


DELETE /tasks/(id)


This command should delete a task.  Try calling this method several times on a URL for a task you have created.  After the first request, GET /tasks/(id) should return a 404.  Subsequent requests should not raise any kind of error.


Setting Up Your Proxy


Once you've verified your new server is working as expected, you have just one step left before you can wire it up to your client.  Currently your new Sinatra server is running on port 4567 while your SproutCore app is on port 4020. This means that by itself your SproutCore app still can't access the Sinatra server.


To solve this, you need to add a "proxy" statement to your SproutCore buildfile.  This will tell SproutCore that certain URLs should be sent to your backend server instead of generating code.


Open the Buildfile in your SproutCore Todos app and add the following line to the bottom:


in todos/Buildfile:

proxy "/tasks", :to => "localhost:4567"


Now any request beginning with /tasks on localhost:4020 will be proxied to your Sinatra backend.  Make sure your Sinatra server is running then restart your sc-server in the SproutCore app and test it out by visiting:




In your web browser.  You should see the JSON you expected to get from Sinatra.  Congratulations, you have a simple web app up and running.  Now you're ready to proceed with writing your Data Source.


Moving On


Go to Step 7: Writing Your Data Source ยป


Comments (8)

Larry Siden said

at 7:11 pm on Oct 27, 2009

tasks.rb from GitHub (commit f077842) doesn't work. When I copied and pasted the code from this page into tasks.rb on my server all was golden. Turns out the problem is "require static_assets". Someone want to fix and push?

Larry Siden said

at 7:21 pm on Oct 27, 2009

Very cool!, but I can't for the life of me figure out where it's putting tasks.db.

puts "sqlite3://#{File.join(File.dirname(__FILE__), 'tmp', 'tasks.db')}"

returns "sqlite3://./tmp/tasks.db", but there is no ./tmp/tasks.db in the dir I am running it tasks.rb from.

Jim Tobin said

at 5:23 pm on Nov 10, 2009

Did you find it? Mine is in /tmp

Larry Siden said

at 7:31 pm on Oct 27, 2009

I'm just learning Ruby. Can someone point me to a URL that will explain why to_json(*a) doesn't result in a stack overflow as it's defined? I'm sure I'm missing something.

Kristian Mandrup said

at 5:13 am on Dec 18, 2009

I am using ruby 1.9.2 and wanted to use mongrel. Can be installed like this:
$ sudo gem install mongrel --source http://gems.rubyinstaller.org

Then just restart sinatra server :) The "default" gem install mongrel fails on 1.9.1 due to some fastthread dependency issues...

Kristian Mandrup said

at 7:49 am on Dec 18, 2009

Aha, I figured out that you have to set the URL to http://localhost:4567/tasks in HTTPClient

Daniel said

at 10:44 am on Jul 9, 2010

When I use POST in HTTP Client it asks for a username and password. What do I enter?

Daniel said

at 8:47 am on Jul 24, 2010

Never mind, I forgot to set the post body. Another question - when I restart my computer, all the tasks are erased because they are in the tmp folder. How can I change where it stores the tasks? I tried changing where it says 'tmp' but I got error messages when I tried to run it in terminal.

You don't have permission to comment on this page.