Angular.js is pretty much my favorite way to develop web applications as of now. For building simple applications with Angular.js I look for a basic backend through which I can add persistence or do some heavy lifting.
Node.js is a one possible backend which does the job and is pretty fast, Rails / Rails-API is another good option. However, being a Scala admirer, I thought of building a backend in Scala. A type safe, compiled language does offer speed and scalability as a backend service.
With Scala as my choice I was left with two options - Play framework or Scalatra. Play framework with Reactive Mongo is a good combination to "play" with but I went for Scalatra as it is lighter and fits the description of a JSON spewing backend well.
With my Scalatra backend, I wanted to use MongoDB for persistent and to finally make things spicier use Akka to make the HTTP calls asynchronous.
After installing and reading the Scalatra documentation, I was on my way. The documentation although good, does leave some gaps, so hope this blog helps someone.
To get the persistence with MongoDB working we need to setup Casbah (the mongo Scala driver). This is simple enough -
libraryDependencies ++= Seq(
"org.mongodb" %% "casbah" % "2.6.0", //add this line in project/build.scala
Similarly, to serialize our response to JSON we add a few dependencies -
libraryDependencies ++= Seq(
"org.mongodb" %% "casbah" % "2.6.0",
"org.scalatra" %% "scalatra-json" % "2.2.1",
"org.json4s" %% "json4s-jackson" % "3.2.4",
Since I am using Scala 2.10.2, there is no need to add any lib for Akka since Akka is built in.
Now, in our Scalatra BootStrap class we setup Akka and Casbah -
import in.rockyj.hello._
import org.scalatra._
import javax.servlet.ServletContext
import com.mongodb.casbah.Imports._
import _root_.akka.actor.{ActorSystem, Props}
class ScalatraBootstrap extends LifeCycle {
val system = ActorSystem("ScalatraSystem")
override def init(context: ServletContext) {
val mongoClient = MongoClient()
val mongoColl = mongoClient("casbah_test")("test_data")
context.mount(new AppServlet(mongoColl, system), "/*")
}
override def destroy(context:ServletContext) {
system.shutdown()
}
}
Simple enough, we setup a casbah_test DB and a simple test_data collection. We also initialize the Akka Actor system and pass these references to a Scalatra servlet.
The Scalatra servlet looks like this -
package in.rockyj.hello
//Scalatra
import org.scalatra._
import scalate.ScalateSupport
import org.scalatra.{FutureSupport, Accepted, ScalatraServlet}
import org.scalatra.json._
import org.scalatra.CorsSupport
// MongoDb-specific imports
import com.mongodb.casbah.Imports._
// JSON-related libraries
import org.json4s.{DefaultFormats, Formats}
//Akka
import _root_.akka.actor.{ActorRef, Actor, Props, ActorSystem}
import _root_.akka.util.Timeout
import _root_.akka.pattern.ask
case class Query(key: String, value: String, mongoColl: MongoCollection)
class AppServlet(mongoColl: MongoCollection, system:ActorSystem) extends MyScalatraWebAppStack with JacksonJsonSupport with FutureSupport with CorsSupport {
val myActor = system.actorOf(Props[MyActor])
protected implicit val timeout = Timeout(10)
protected implicit val jsonFormats: Formats = DefaultFormats
protected implicit def executor = system.dispatcher
options("/*") {
response.setHeader("Access-Control-Allow-Headers", request.getHeader("Access-Control-Request-Headers"));
}
get("/") {
contentType="text/html"
layoutTemplate("/WEB-INF/templates/views/index.jade")
}
/**
* Insert a new object into the database:
*curl -i -H "Accept: application/json" -X POST -d "key=hello&value=world" http://localhost:8080/insert
*/
post("/insert") {
val key = params("key")
val value = params("value")
val newObj = MongoDBObject(key -> value)
mongoColl += newObj
}
/**
* Query for the first object which matches the values given.
* try http://localhost:8080/query/hello/world in your browser.
*/
get("/query/:key/:value") {
contentType = formats("json")
val q = Query(params("key"), params("value"), mongoColl)
myActor ? q
}
}
class MyActor extends Actor {
def receive = {
case Query(key, value, mongoColl) => {
val q = MongoDBObject(key -> value)
val res = mongoColl.findOne(q).getOrElse(Map("not found" -> true))
sender ! res
}
}
}
Nothing really complicated, we can seed some data using CURL at the available POST end point. Since we will be calling these services from a Angular.js app running on a different server I needed to add CORS support via this line -
options("/*") {
response.setHeader("Access-Control-Allow-Headers", request.getHeader("Access-Control-Request-Headers"));
}
The whole Scalatra app is available on github.
On the Angular.js side, we will consume this service and display the data on a web page.
#app.coffee
'use strict'
angular.module('demoApp', ["demoApp.webService"])
.config ($routeProvider) ->
$routeProvider
.when('/', {templateUrl: 'views/main.html', controller: 'MainCtrl'})
.otherwise(redirectTo: '/')
#webservice.coffee
"use strict"
class WebService
constructor: (@$http) ->
getGreeting: () ->
promise = @$http.get("http://localhost:8080/query/hello/world")
angular.module "demoApp.webService", [], ($provide) ->
$provide.factory "webService", ["$http", ($http) -> new WebService($http)]
#main.coffee
'use strict'
class MainCtrl
constructor: (@$scope, @webService) ->
promise = @webService.getGreeting()
promise.then @success, @error
success: (response) =>
@$scope.response = response.data
error: (response) =>
#do something
MainCtrl.$inject = ["$scope", "webService"]
angular.module("demoApp").controller "MainCtrl", MainCtrl
That is it for this weekend's hack. I ran some tests with Apache benchmark on my Scalatra server in a single core VM and the average response for - ab -n 200 -c 4 http://localhost:8080/query/hello/world came to be 14ms which is pretty good.