path: root/railties/guides/source/api_app.textile

h2. Using Rails for API-only Apps

In this guide you will learn:

* What Rails provides for API-only applications
* How to configure Rails to start without any browser features
* How to decide which middlewares you will want to include
* How to decide which modules to use in your controller

NOTE: This guide reflects features that have not yet been fully implemented. Docs first :)

endprologue.

h3. What is an API app?

Traditionally, when people said that they used Rails as an "API", they meant providing a programmatically accessible API alongside their web application.
For example, GitHub provides "an API":http://developer.github.com that you can use from your own custom clients.

With the advent of client-side frameworks, more developers are using Rails to build a backend that is shared between their web application and other native applications.

For example, Twitter uses its "public API":https://dev.twitter.com in its web application, which is built as a static site that consumes JSON resources.

Instead of using Rails to generate dynamic HTML that will communicate with the server through forms and links, many developers are treating their web application as just another client, delivered as static HTML, CSS and JavaScript, and consuming a simple JSON API

This guide covers building a Rails application that serves JSON resources to an API client *or* client-side framework.

h3. Why use Rails for JSON APIs?

The first question a lot of people have when thinking about building a JSON API using Rails is: "isn't using Rails to spit out some JSON overkill? Shouldn't I just use something like Sinatra?"

For very simple APIs, this may be true. However, even in very HTML-heavy applications, most of an application's logic is actually outside of the view layer.

The reason most people use Rails is that it provides a set of defaults that allows us to get up and running quickly without having to make a lot of trivial decisions.

Let's take a look at some of the things that Rails provides out of the box that are still applicable to API applications.

Handled at the middleware layer:

* Reloading: Rails applications support transparent reloading. This works even if your application gets big and restarting the server for every request becomes non-viable.
* Development Mode: Rails application come with smart defaults for development, making development pleasant without compromising production-time performance.
* Test Mode: Ditto test mode.
* Logging: Rails applications log every request, with a level of verbosity appropriate for the current mode. Rails logs in development include information about the request environment, database queries, and basic performance information.
* Security: Rails detects and thwarts "IP spoofing attacks":http://en.wikipedia.org/wiki/IP_address_spoofing and handles cryptographic signatures in a "timing attack":http://en.wikipedia.org/wiki/Timing_attack aware way. Don't know what an IP spoofing attack or a timing attack is? Exactly.
* Parameter Parsing: Want to specify your parameters as JSON instead of as a URL-encoded String? No problem. Rails will decode the JSON for you and make it available in +params+. Want to use nested URL-encoded params? That works too.
* Conditional GETs: Rails handles conditional +GET+, (+ETag+ and +Last-Modified+), processing request headers and returning the correct response headers and status code. All you need to do is use the "stale?":http://api.rubyonrails.org/classes/ActionController/ConditionalGet.html#method-i-stale-3F check in your controller, and Rails will handle all of the HTTP details for you.
* Caching: If you use +dirty?+ with public cache control, Rails will automatically cache your responses. You can easily configure the cache store.
* HEAD requests: Rails will transparently convert +HEAD+ requests into +GET+ requests, and return just the headers on the way out. This makes +HEAD+ work reliably in all Rails APIs.

While you could obviously build these up in terms of existing Rack middlewares, I think this list demonstrates that the default Rails middleware stack provides a lot of value, even if you're "just generating JSON".

Handled at the ActionPack layer:

* Resourceful Routing: If you're building a RESTful JSON API, you want to be using the Rails router. Clean and conventional mapping from HTTP to controllers means not having to spend time thinking about how to model your API in terms of HTTP.
* URL Generation: The flip side of routing is URL generation. A good API based on HTTP includes URLs (see "the GitHub gist API":http://developer.github.com/v3/gists/ for an example).
* Header and Redirection Responses: +head :no_content+ and +redirect_to user_url(current_user)+ come in handy. Sure, you could manually add the response headers, but why?
* Content Negotiation: The Rails +respond_to+ and +respond_with+ features automatically figure out which MIME type to serve, based on the request's +Accept+ header and available types. If you ever need to add support for types other than JSON (XML, CSV, or some proprietary format), this will come in handy.
* Caching: Rails provides page, action and fragment caching. Fragment caching is especially helpful when building up a nested JSON object.
* Basic, Digest and Token Authentication: Rails comes with out-of-the-box support for three kinds of HTTP authentication.
* Instrumentation: Rails 3.0 added an instrumentation API that will trigger registered handlers for a variety of events, such as action processing, sending a file or data, redirection, and database queries. The payload of each event comes with relevant information (for the action processing event, the payload includes the controller, action, params, request format, request method and the request's full path).
* Generators: This may be passé for advanced Rails users, but it can be nice to generate a resource and get your model, controller, test stubs, and routes created for you in a single command.
* Plugins: Many third-party libraries come with support for Rails that reduces or eliminates the cost of setting up and gluing together the library and the web framework. This includes things like overriding default generators, adding rake tasks, and honoring Rails choices (like the logger and cache backend).

Of course, the Rails boot process also glues together all registered components. For example, the Rails boot process is what uses your +config/database.yml+ file when configuring ActiveRecord.

**The short version is**: you may not have thought about which parts of Rails are still applicable even if you remove the view layer, but the answer turns out to be "most of it".

h3. The Basic Configuration

If you're building a Rails application that will be an API server first and foremost, you can start with a more limited subset of Rails and add in features as needed.

You can generate a new bare Rails app:

<shell>
$ rails new my_api --http
</shell>

This will do three main things for you:

* Configure your application to start with a more limited set of middleware than normal. Specifically, it will not include any middleware primarily useful for browser applications (like cookie support) by default.
* Make +ApplicationController+ inherit from +ActionController::HTTP+ instead of +ActionController::Base+. As with middleware, this will leave out any +ActionController+ modules that provide functionality primarily used by browser applications.
* Configure the generators to skip generating views, helpers and assets when you generate a new resource.

If you want to take an existing app and make it an API app, follow the following steps.

In +config/application.rb+ add the following lines at the top of the +Application+ class:

<ruby>
config.middleware.http_only!
config.generators.http_only!
</ruby>

Change +app/controllers/application_controller.rb+:

<ruby>
# instead of
class ApplicationController < ActionController::Base
end

# do
class ApplicationController < ActionController::HTTP
end
</ruby>

h3. Choosing Middlewares

An API application comes with the following middlewares by default.

* +Rack::Cache+: Caches responses with public +Cache-Control+ headers using HTTP caching semantics. See below for more information.
* +Rack::Sendfile+: Uses a front-end server's file serving support from your Rails application.
* +Rack::Lock+: If your application is not marked as threadsafe (+config.threadsafe!+), this middleware will add a mutex around your requests.
* +ActionDispatch::RequestId+:
* +Rails::Rack::Logger+:
* +ActionDispatch::ShowExceptions+: Rescue exceptions and re-dispatch them to an exception handling application
* +ActionDispatch::DebugExceptions+: Log exceptions
* +ActionDispatch::RemoteIp+: Protect against IP spoofing attacks
* +ActionDispatch::Reloader+: In development mode, support code reloading.
* +ActionDispatch::ParamsParser+: Parse XML, YAML and JSON parameters when the request's +Content-Type+ is one of those.
* +ActionDispatch::Head+: Dispatch +HEAD+ requests as +GET+ requests, and return only the status code and headers.
* +Rack::ConditionalGet+: Supports the +stale?+ feature in Rails controllers.
* +Rack::ETag+: Automatically set an +ETag+ on all string responses. This means that if the same response is returned from a controller for the same URL, the server will return a +304 Not Modified+, even if no additional caching steps are taken. This is primarily a client-side optimization; it reduces bandwidth costs but not server processing time.

Other plugins, including +ActiveRecord+, may add additional middlewares. In general, these middlewares are agnostic to the type of app you are building, and make sense in an API-only Rails application.

You can get a list of all middlewares in your application via:

<shell>
$ rake middleware
</shell>

h4. Using Rack::Cache

When used with Rails, +Rack::Cache+ uses the Rails cache store for its entity and meta stores. This means that if you use memcache, for your Rails app, for instance, the built-in HTTP cache will use memcache.

To make use of +Rack::Cache+, you will want to use +stale?+ in your controller. Here's an example of +stale?+ in use.

<ruby>
def show
  @post = Post.find(params[:id])

  if stale?(:last_modified => @post.updated_at)
    render json: @post
  end
end
</ruby>

The call to +stale?+ will compare the +If-Modified-Since+ header in the request with +@post.updated_at+. If the header is newer than the last modified, this action will return a +304 Not Modified+ response. Otherwise, it will render the response and include a +Last-Modified+ header with the response.

Normally, this mechanism is used on a per-client basis. +Rack::Cache+ allows us to share this caching mechanism across clients. We can enable cross-client caching in the call to +stale?+

<ruby>
def show
  @post = Post.find(params[:id])

  if stale?(:last_modified => @post.updated_at, :public => true)
    render json: @post
  end
end
</ruby>

This means that +Rack::Cache+ will store off +Last-Modified+ value for a URL in the Rails cache, and add an +If-Modified-Since+ header to any subsequent inbound requests for the same URL.

Think of it as page caching using HTTP semantics.

NOTE: The +Rack::Cache+ middleware is always outside of the +Rack::Lock+ mutex, even in single-threaded apps.

h4. Using Rack::Sendfile

When you use the +send_file+ method in a Rails controller, it sets the +X-Sendfile+ header. +Rack::Sendfile+ is responsible for actually sending the file.

If your front-end server supports accelerated file sending, +Rack::Sendfile+ will offload the actual file sending work to the front-end server.

You can configure the name of the header that your front-end server uses for this purposes using +config.action_dispatch.x_sendfile_header+ in the appropriate environment config file.

You can learn more about how to use +Rack::Sendfile+ with popular front-ends in "the Rack::Sendfile documentation":http://rubydoc.info/github/rack/rack/master/Rack/Sendfile

The values for popular servers once they are configured to support accelerated file sending:

<ruby>
# Apache and lighttpd
config.action_dispatch.x_sendfile_header = "X-Sendfile"

# nginx
config.action_dispatch.x_sendfile_header = "X-Accel-Redirect"
</ruby>

Make sure to configure your server to support these options following the instructions in the +Rack::Sendfile+ documentation.

NOTE: The +Rack::Sendfile+ middleware is always outside of the +Rack::Lock+ mutex, even in single-threaded apps.

h4. Using ActionDispatch::ParamsParser

+ActionDispatch::ParamsParser+ will take parameters from the client in JSON and make them available in your controller as +params+.

To use this, your client will need to make a request with JSON-encoded parameters and specify the +Content-Type+ as +application/json+.

Here's an example in jQuery:

<plain>
jQuery.ajax({
  type: 'POST',
  url: '/people'
  dataType: 'json',
  contentType: 'application/json',
  data: JSON.stringify({ person: { firstName: "Yehuda", lastName: "Katz" } }),

  success: function(json) { }
});
</plain>

+ActionDispatch::ParamsParser+ will see the +Content-Type+ and your params will be +{ :person => { :firstName => "Yehuda", :lastName => "Katz" } }+.

h4. Other Middlewares

Rails ships with a number of other middlewares that you might want to use in an API app, especially if one of your API clients is the browser:

* +Rack::SSL+: Redirects any HTTP request to HTTPS.
* +Rack::Runtime+: Adds a header to the response listing the total runtime of the request.
* +Rack::MethodOverride+: Allows the use of the +_method+ hack to route POST requests to other verbs.
* +ActionDispatch::Cookies+: Supports the +cookie+ method in +ActionController+, including support for signed and encrypted cookies.
* +ActionDispatch::Flash+: Supports the +flash+ mechanism in +ActionController+.
* +ActionDispatch::BestStandards+: Tells Internet Explorer to use the most standards-compliant available renderer. In production mode, if ChromeFrame is available, use ChromeFrame.
* Session Management: If a +config.session_store+ is supplied, this middleware makes the session available as the +session+ method in +ActionController+.

Any of these middlewares can be adding via:

<ruby>
config.middleware.use Rack::MethodOverride
</ruby>

h4. Removing Middlewares

If you don't want to use a middleware that is included by default in the API-only middleware set, you can remove it using +config.middleware.delete+:

<ruby>
config.middleware.delete ::Rack::Sendfile
</ruby>

Keep in mind that removing these features may remove support for certain features in +ActionController+.

h3. Choosing Controller Modules

An API application (using +ActionController::HTTP+) comes with the following controller modules by default:

* +AbstractController::Translation+: Support for the +l+ and +t+ localization and translation methods. These delegate to +I18n.translate+ and +I18n.localize+.
* +ActionController::UrlFor+: Makes +url_for+ and friends available.
* +ActionController::Redirecting+: Support for +redirect_to+
* +ActionController::Renderers::JSON+: Support for +render :json+
* +ActionController::ConditionalGet+: Support for +stale?+
* +ActionController::RackDelegation+: Support for the +request+ and +response+ methods returning +ActionDispatch::Request+ and +ActionDispatch::Response+ objects.
* +ActionController::MimeResponds+: Support for content negotiation (+respond_to+, +respond_with+)
* +ActionController::DataStreaming+: Support for +send_file+ and +send_data+
* +AbstractController::Callbacks+: Support for +before_filter+ and friends
* +ActionController::Instrumentation+: Support for the instrumentation hooks defined by +ActionController+ (see "the source":https://github.com/rails/rails/blob/master/actionpack/lib/action_controller/metal/instrumentation.rb for more).

Other plugins may add additional modules. You can get a list of all modules included into +ActionController::HTTP+ in the rails console:

<shell>
$ irb
>> ActionController::HTTP.ancestors - ActionController::Metal.ancestors
</shell>

h4. Adding Other Modules

All ActionController modules know about their dependent modules, so you can feel free to include any modules into your controllers, and all dependencies will be included and set up as well.

Some common modules you might want to add:

* +ActionController::HTTPAuthentication::Basic+ (or +Digest+ or +Token): Support for basic, digest or token HTTP authentication.
* +ActionController::Rendering+: Support for templating and +ActionView+.
* +AbstractController::Layouts+: Support for layouts when rendering.
* +ActionController::Renderers::XML+: Support for +render :xml+.
* +ActionController::Cookies+: Support for +cookies+, which includes support for signed and encrypted cookies. This requires the cookie middleware.
* +ActionController::Rescue+: Support for +rescue_from+.

The best place to add a module is in your +ApplicationController+. You can also add modules to individual controllers.