Poroute

Poroute is a Rack router for defining controllers using Plain Old Ruby Objects (POROs).

Features

• Controllers are defined using POROs
• Actions (route handlers) accept Hashes and return Hashes
• Routes are defined using an unambiguous DSL
• Middlewares use continuation passing style

Usage

Define your controller:

# app/controllers/posts.rb

module Controllers
  class Posts
    # For this example, we're using an in-memory Hash to store state.
    # In a real application, we'd probably be reaching out to a database.
    def initialize
      @posts = {}
      @next_id = 0
    end

    def create(request)
      post = request.dig(:body, :post)
      @posts[@next_id] = post
      @next_id += 1

      {
        status: 201,
        headers: { 'X-Total-Posts' => @posts.length },
        body: post
      }
    end

    def read(params:)
      id = params[:id].to_i

      if (post = @posts[id])
        {
          status: 200,
          body: post
        }
      else
        {
          status: 404,
          body: { error: "Unable to find post with ID #{id}" }
        }
      end
    end
  end
end

Define your routes:

# app/routes.rb

require 'poroute'
require 'controllers/posts'

Routes = Poroute.define do
  scope '/api/posts' do
    controller Controllers::Posts

    middleware Poroute::Middleware::Json

    post '/', :create
    get '/:id', :read
  end
end

Start your application:

# config.ru

$LOAD_PATH << File.expand_path('app', __dir__)
require 'routes'

run Routes

Controllers & Actions

Controllers are standard Ruby objects -- no need to subclass, include, or extend anything. Unlike Rails and Sinatra, where controllers are instantiated once per request, Poroute controllers are instantiated once and for all, so be careful with memoization. Actions are methods defined on a controller which accept an HTTP request and return an HTTP response, both represented by Hashes.

In the above example, the Controllers::Posts#create action accepts the request as a Hash, appends the post to the data store, and returns a 201. Many actions, such as Controllers::Posts#read, only need a subset of the data from the request.

To make things easier, Poroute allows users to define actions using keyword arguments to select the desired request keys. The following keyword arguments can be required by actions:

• method - HTTP method
• path - HTTP path (before normalization)
• query - query parameters (as a String)
• headers - Hash containing the HTTP headers (without the HTTP_ prefix added by Rack)
• body - request body (as an IO-like object)
• params - Hash of param parsed from the path description
• env - Rack env

Actions must return a Hash which includes a :status, and optional :headers and :body keys.

Routes

The routing DSL maps descriptions of HTTP requests to controllers, actions, and middlewares. Unlike the Rails routing DSL, Poroute forces users to specify which controller will be hit for each scope. Forcing the specification is intended to remove ambiguity from the routes:action mappings.

Controllers are specified with the controller DSL method, which can either accept a controller class, or an instance of a controller class. If a class is specified, it must have a 0 argument initializer.

Routes may be declared using head, get, post, put, patch, and delete. Each of these methods accepts two arguments: a path pattern and a Symbol which corresponds to one of the controller's actions.

Path patterns can match HTTP paths by exact String literal, such as by specifying get '/about', :about. Path segments (i.e. the parts of the path between /s) can also be used to match any String, and bind that value to a variable. For instance, a match against /users/:user_id/posts/:posts_id with the path /users/tom/posts/poros would bind { user_id: 'tom', post_id: 'poros' }.

To (greedily) match zero or more path segments, use * instead of :. A match against /start/*middle/finish with the path /start/a/b/c/d/e/finish would bind { middle: 'a/b/c/d/e' }. Generally, this is used to define 404 handlers.

A set of route:action mappings may be mounted under a scope. Like the route mapping methods, scope accepts a path pattern, which follows the same rules described above. Each scope inherits the middleware from its parent scope. As of right now, only one controller may be used per scope. If you need to mount multiple controllers under a common prefix, you can define multiple scopes with that prefix.

Middleware

Poroute middleware is defined in Continuation-passing style. Essentially, each layer of middleware is a function that accepts a request. To pass the request, or a modified version of it, onto the next middleware in the stack, yield it. Doing so will return the result from that layer of the middleware stack. This result can then either be returned or modified.

Here's an example:

class MarketingMiddleware
  def call(request)
    response = yield request

    response.merge(
      headers: (response[:headers] || {}).merge(
        'X-Routed-With' => "Poroute v#{Poroute::VERSION}"
      )
    )
  end
end

When mounted, MarketingMiddleware will add an X-Routed-With header to every HTTP response.

Motivation

Poroute is far from the first routing library for Ruby/Rack. Its routing DSL even heavily resembles that of Rails (though with the caveat that controllers must be specified, which is a feature). So, what's it add?

Great question, my fine straw man. However, the main feature of Poroute is what it doesn't add. Instead of using a DSL to define controller actions, they're just normal methods on normal classes.

Want to see how an action would behave with different arguments? Call it!

Want to delegate a request to a different action? Call it!

Do you need a recursive route handler? Call i--wait, no, probably don't do that.

Joking aside, methods are easier to understand, easier to test, and grant the developer at least as much freedom as a DSL.

This is also why actions are simple methods which accept requests and return responses. There's no ambiguity about how to acheive what you want ("How do I set the response status? Do I return it? Is there some status DSL method?")

In short, Poroute is designed to be intuitive and unobtrusive in a way that other routing libraries are not.

Installation

Add this line to your application"s Gemfile:

gem 'poroute'

And then execute:

$ bundle install

Or install it yourself as:

$ gem install poroute

Development

After checking out the repo, run bin/setup to install dependencies. Then, run bundle exec rake to run the tests and code quality metrics. You can also run bin/console for an interactive prompt that will allow you to experiment.

To install this gem onto your local machine, run bundle exec rake install. To release a new version, update the version number in version.rb, and then run bundle exec rake release, which will create a git tag for the version, push git commits and tags, and push the .gem file to rubygems.org.

Contributing

Bug reports and pull requests are welcome on GitHub at https://github.com/nahiluhmot/poroute. This project is intended to be a safe, welcoming space for collaboration, and contributors are expected to adhere to the code of conduct.

License

The gem is available as open source under the terms of the MIT License.

Code of Conduct

Everyone interacting in the Poroute project's codebases, issue trackers, chat rooms and mailing lists is expected to follow the code of conduct.