Sleep

Rest server framework, but not sinatra like, but more object (resource) oriented, as it should be!

Altough it is written in Coffeescript (and so are all examples), it is compiled and can be used in every javascript project.

Quickstart

$ npm install sleeprest

And then get it a simple hello world example using

Sleep = require 'sleeprest'
 
# Create a new server, and create a new resource 
server = new Sleep
hello = server.resource 'hello'
 
# Add a handler to it to set a header 
hello.use (req) ->
  @header 'x-hello', 'yes'
 
# Create a subresource with get handler 
hello.resource('world').get (req) ->
  message: "Hello world!"
 
# Another but this one is variable 
hello.resource(':name').get (req) ->
  {name} = req.params # Just like restify 
  message: "Hello #{name}.."
 
# Let the server listen 
server.listen 1337
 

API

Resource

In Sleep, everything is a resource. This is the main class you get when doing require 'sleeprest'.

new Resource([options])

This creates a resource with nothing on him. Most of the time you want to start this way, and add subresources from here.

You may provide an options object, more on that soon.

.get|post|put|delete([Function<request> before...,] Function<request> handler)

Add a function to handle request to the current resource, with the specified method. This will not catch anything flowing down to a subresource. the before functions will run before the handler, so they can alter the request and response, or throw an error.

The return value of the handler function is converted into the desired format and send as response. You can use promises to make you code asynchronous. Personally I use and recommend bluebird which is to be found here: https://github.com/petkaantonov/bluebird

When an array is returned it is wrapped

.use(Function<request>)

Add a function to the general handler chain. It will run on every request and will be able to alter the request and response. Connect/Express like.

.resource(path[, names])

Alias: .res(path[, name])

Create a subresource. This will catch anything that doesn't fit in the current resource, and matches the path description you've given.

Path can be a string or a RegExp. When it is a string, it is parsed sinatra like: a path with possible :variable values. A very common pattern is

parent = server.resource 'items'
### ... add code to list and post items ... ###
 
child = parent.resource ':item'
### ... add code to get, put and delete the item ... ###

When a variable like :item is found, the value will be stored in req.params with the variable name as key (eg: item: value).

When requesting the url GET /items/5 it will run the handler on the child resource with req.params = item: 5.

If the path is a RexExp, it will be matched and if it matches, it will be stored in req.params too, with the capture groups as values and names as keys. When names is not an array, it will be wrapped in one :p

.set(key[, value])

Sets an option value for the resource. Only two options right now:

• hal: Whether or not to enable HAL support.
• debug: Whether or not to show much information.

.listen(port, ...)

Does exactly the same as the http server listen method. And yes, every resource can be turned into a server, even while it also used as a subresource! http://nodejs.org/api/http.html#http_server_listen_port_hostname_backlog_callback

Request

http://nodejs.org/api/http.html#http_http_createserver_requestlistener Only has more properties set by handlers and has .params filled with the url params.

Response

Soon... ^^