Columbo

RIMMER: You can scoff, Lister. That's nothing new. They laughed at Galileo. They laughed at Edison. They laughed at Columbo.
LISTER: Who's Columbo?
RIMMER: The man with the dirty mac who discovered America.

Red Dwarf Season 1, Episode 4

Discovers REST resources in js files named by convention. Is it a good idea? I don't know. Either way, it makes setting up a REST interface pretty easy.

Usage

URLs

Put your resource classes in a folder (the default is name is resources). Follow the convention:

Library.js
LibraryBook.js

or

LibraryResource.js
LibraryBookResource.js

This will generate resources for the following URLs:

/libraries
/libraries/{libraryId}/books
/libraries/{libraryId}/books/{bookId}

Methods

HTTP Methods are determined by the methods exposed on your resource classes. An OPTIONS resource is generated dynamically.

Your methods should accept whatever arguments your web framework requires.

LibraryBook = function() {
 
};
 
LibraryBook.prototype.retrieve = function() {
  // generates GET with a URL parameter named 'id'
};
 
LibraryBook.prototype.retrieveAll = function() {
  // generates GET with no URL parameter
};
 
LibraryBook.prototype.retrieveOne = function() {
  // generates GET with no URL parameter (intended for singletons, do no use with retrieve and retrieveAll)
};
 
LibraryBook.prototype.create = function() {
  // generates POST
};
 
LibraryBook.prototype.update = function() {
  // generates PUT
};
 
LibraryBook.prototype.patch = function() {
  // generates PATCH
};
 
LibraryBook.prototype.remove = function() {
  // generates DELETE
};

API

var Columbo = require("columbo");
 
// all options are, well, optional
var columbo = new Columbo({
  // The path to the resources directory
  resourceDirectory: "./resources",
 
  // All files in `resourceDirectory` will be `require`d - pass a function here
  // to override how the resources are created.  `resource` will be whatever is
  // returned from `require` and `name` is the camelised name of the file
  resourceCreator: function(resource, name) {
    return new resource()
  },
 
  // This will be called for every resource found - if you need to edit the resource
  // definition, do it here then pass it to the callback
  preProcessor: function(resource, callback) {
    callback(undefined, resource)
  },
 
  // In order to support :id or {id} form url arguments, pass a function here
  // that will return arguments in the format you desire.
  idFormatter: function(id) {
    return "{" + id + "}";
  },
 
  // Responder for OPTIONS requests - the first argument will be an array of
  // strings corresponding to HTTP verbs, subsequent arguments are supplied by
  // your web framework of choice.
  optionsSender: function(opts, request) {
    request.reply(opts);
  },
 
  // Optionally pass in a logger (e.g. Winston) - defaults to console.
  logger: {}
});
 
// Discover resources
var resources = columbo.discover();

Above resources will be an array:

[
  {
    method: String,		// e.g. GET
    path: String,		// e.g. /foo
    handler: Function	// a bound function - the method on the resource class
  }, ...
]

You can then use it with Hapi

var columbo = new Columbo();
 
var server = new Hapi.Server();
server.addConnection({port: 3000});
 
columbo.discover(function(error, routes) {
  server.route(routes);
  server.start();
})

Or with Express:

var app = express();
...
var columbo = new Columbo({
  idFormatter: function(id) {
    return ":" + id;
  },
  optionsSender: function(options, request, response) {
    response.json(options);
  }
});
columbo.discover(function(error, routes) {
  routes.forEach(function(resource) {
    app[resource.method.toLowerCase()](resource.path, resource.handler);
  });
})

Etc, etc.

Todo

• Framework agnostic argument validation
• OPTIONS is a little half baked