actionhero Tutorial
- created: June 22, 2013
- updated: August 16, 2014
This guide will walk you through the creation of the application in this repository, and in the process you will learn some of the basics of actionhero.
You will become comfortable with the following topics:
A simple Blogging API & site:
- Getting Started with a new actionhero Project
- Creating Initializers
- Users & Authenticationn
- Public and Private actions with Middleware
- Creating Actions
- Routes
- Testing
- Consuming the API via the Web
Adding a chat room
Creating a Custom Server
Next Steps
actionhero Resources
Notes
- You are expected to have a basic familiarity with node.js and the command line
- You do not need to clone this repository to follow this guide. The code in this repository represents the final state of a project created with these instructions. The code in this project should serve as a reference.
- You should also be able to run this project by:
git clone https://github.com/evantahler/actionhero-tutorial.git
cd actionhero-tutorial
npm install
npm start
- There are a few extra steps needed to persists data to redis and to use the twitter server example discussed below.
- This project uses redis as a database. actionhero comes with fakeRedis, which is an in-process redis server, but it does not persist data. If you want to use this project in a cluster or across multiple servers, you need to install and use a real redis server. Change the appropriate
redis
sections in/config/redis.js
to enable this. - Remember that actionhero is an API server, so we will be focusing on creating an API for blogging and chatting, and applying that to a website rather than creating a beautiful website itself.
Getting Started with a new actionhero Project
files discussed in this section:
relevant documentation section:
actionhero is a node.js package. Be sure you have node.js (version >= 8.0.0) installed. Node now also comes with npm, the node package manager. You can get node from nodejs.org if you do not have it.
This guide was written on OSX 10.8 It should be appropriate for any version of OSX > 10.6. It should also work on most Linux distributions (Ubuntu, CentOs, Fedora, etc). The concepts presented here should also be appropriate for windows users, but many of the "Getting Started" commands will not work as described here. If you are looking for help on getting started on Windows, Mark Tucker has a video tutorial for windows users
Create a new directory for this project and enter it (in the terminal):
mkdir ~/actionhero-tutorial
cd ~/actionhero-tutorial
note: From this point forward, it is assumed that all commands listed are run from within the ~/actionhero-tutorial
directory.
Install the actionhero package locally from NPM:
npm install actionhero
Use the actionhero generator to build your project
./node_modules/.bin/actionhero generate
Install any project dependancies
npm install
Try to boot the actionhero server
npm start
You should see the default actionhero welcome page at http://localhost:8080/public
(visit in your browser)
The port 8080
is defined in /config/servers/web.js
, along with all other settings for actionhero. actionhero has 2 types of http routes: static files and api routes. static files are served from /public
and the api is served from /api
. These routes are configurable. actionhero also picks one of these to be the default root route. This is defined by api.config.servers.web.rootEndpointType
. As we want to make a website, lets change that from api
to file
.
Restart your server by pressing ctrl+c
in the terminal window running actionhero. Start up the server again and visit http://localhost:8080/
and you should see the welcome page. You will note that the setting we just changed was under the servers.web
section. This is because this setting is only relevant to HTTP clients, and not the others (socket, websocket, etc). We will talk about these more later.
We should also enable all the servers which ship with actionhero (web, websocket, and socket). Enable their sections in thier config files
Lets change one more thing in config/api.js
: development mode. Change api.config.general.developmentMode = true;
Development mode is helpful while creating a new application as it will automatically restart your server on configuration changes, and watch and reload your actions and tasks as you change them.
actionhero uses the variable NODE_ENV
to determine which modification file to load from /config/*
to load to modify the default values in api.config
. This is how you can set different variables per envrionment. We will use this for testing later.
Creating Initializers
files discussed in this section:
relevant documentation section:
initializers in actionhero are places to save common code which other parts of your application will use. Here is where you might connect to your database or define middlewares. Normally, you append your new classes to the api
object so it becomes available in scope for your actions and tasks (like api.mysql
or api.game
). When your server boots, you can optionally execute code within a _start
method.
Because we are building a blog, first we will need a place to store our entries and comments. As actionhero already has redis available under api.redis.client
, lets use that for our data store.
Create a new initializer for the blog's common functions:
./node_modules/.bin/actionhero generateInitializer --name=blog
I'll define some common blog functions we'll use later in actions in initializers/blog.js
Our blogging methods are:
api.blog = {
// constants
separator: ";",
postPrefix: "posts",
commentPrefix: "comments:",
// posts
postAdd: function(userName, title, content, next){},
postView: function(userName, title, next){},
postsList: function(userName, next){},
postEdit: function(userName, title, content, next){},
postDelete: function(userName, title, next){},
// comments
commentAdd: function(userName, title, commenterName, comment, next){},
commentsView: function(userName, title, next){},
commentDelete: function(userName, title, commentId, next){},
}
A few things to note:
- posts are hashes with the content and some additional meta data
- comments are also hashes, a key for each comment
- we always make asynchronous functions, and they always return
callback(error, data)
, a common node.js pattern. - at this layer, we don't worry about authentication or validations
Users & Authentication
files discussed in this section:
We know we will need to authenticate users to our blog, so lets make another initializer to handle this as well.
./node_modules/.bin/actionhero generateInitializer --name=users
Our user methods are:
api.users = {
// constants
usersHash: "users",
// methods
add: function(userName, password, next){},
list: function(next){},
authenticate: function(userName, password, next){},
delete: function(userName, password, next){},
}
Notes:
- We are again storing all data in a redis hash
- If we delete a user, we should delete all the posts and comments from them
- We are only using md5 as a hashing algorithm for our user's passwords. There are more secure options which you should use in production, like BCrypt
Public and Private actions with Middleware
files discussed in this section:
relevant documentation section:
In the steps above, we created a api.users.authenticate
method, but didn't use it anywhere. There are clearly methods which should be protected (like adding a post, or deleting a user), but we need to safeguard them somehow.
In actionhero, we know that we will wrap any use of our initializer's methods by users in actions
, so we can create a middleware which we can apply to these actions
.
Let's create a new initializer for this:
./node_modules/.bin/actionhero generateInitializer --name=middleware
There are arrays of functions in actionhero which will be run before an after every action. Here, we only need a check before to see if an action should be run. You have access to the action itself, along with the connection.
The middleware we created allows us to simply append action.authenticated = true
to the action defintion, and the middleware will be invoked.
Middlewares are invoked by adding them to the api.actions.addPreProcessor(authenticationMiddleware);
collection
Creating Actions
files discussed in this section:
relevant documentation section:
Now that we have our helpers for getting and setting blog posts, how can we allow users to use them? Actions!
Action files can define a few actions each, so lets create one for comments and one for posts.
./node_modules/.bin/actionhero generateAction --name=users
./node_modules/.bin/actionhero generateAction --name=blog
Note how we added action.authenticated = true
on the actions which required security.
Now we can use CURL to test out our API! Note that right now, all HTTP methods will work (get, post, etc). We'll be setting up routing next. Be sure to URL-encode all your input.
- Add a user:
curl -X POST -d "userName=evan" -d "password=password" "http://localhost:8080/api/userAdd"
- Check that you can log in:
curl -X POST -d "userName=evan" -d "password=password" "http://localhost:8080/api/authenticate"
- Add a post:
curl -X POST -d "userName=evan" -d "password=password" -d "title=first-post" -d "content=My%20first%20post.%20%20Yay." "http://localhost:8080/api/postAdd"
- View the post
curl -X POST -d "userName=evan" -d "title=first-post" "http://localhost:8080/api/postView"
- Add a comment
curl -X POST -d "userName=evan" -d "title=first-post" -d "comment=cool%20post" -d "commenterName=someoneElse" "http://localhost:8080/api/commentAdd"
- View the comments
curl -X POST -d "userName=evan" -d "title=first-post" "http://localhost:8080/api/commentsView"
Routes
files discussed in this section:
relevant documentation section:
We have the basics of our API working, but it might be tedious to keep using GET and POST params. It's time to set up routes. Routes allow different HTTP verbs to preform a different action on the same URL. We'll use a routes.js
file to transform our API into restful resources for users, comments, and posts. You can derive input variables from the structure of URLs with routing as well.
Now we can get the list of posts for user evan
with curl -X GET "http://localhost:8080/api/posts/evan"
and we don't need to pass any parameters.
- I can get a list of my posts with
http://localhost:8080/api/posts/evan
- I can view a post with
http://localhost:8080/api/post/evan/first-post
- etc
Testing
files discussed in this section:
There are many testing tools and packages which exist for nodejs. actionhero is not opinionated about which testing framework you should use, but nonetheless, testing is important! actionhero exposes a number of utilities to make it easy to boot up a server with configuration overrides to make testing easier.
Lets setup a test with the mocha
and should
packages. We'll use the request
package to make HTTP requests simpler in our tests.
npm install mocha should request
(and add them to your package.json
in the devDependencies
section).
First, lets create a spec helper in test/_setup.js
Now we can use our _setup.js
in a test. Let's create an integration test /test/integration.js
for post creation and reading. Note how we are using fakeredis
so we have an isolated in-process test database to work with.
We can now run the test with the mocha
command. In our package.json
we can also setup npm test
to run the test suite how we would like it: "test": "NODE_ENV=test ./node_modules/.bin/mocha --reporter spec ./test"
Note how we are calling NODE_ENV=test
to tell actionhero we are running this command in the 'test' environment this will signal actionhero load any configChanges from the /config/environments/test.js
file. Here we setup redis and the servers how we want for testing.
A successful test run looks like this:
Consuming the API via the Web
files discussed in this section:
relevant documentation section:
actionhero is primarily an API server, but it can still serve static files for you. In config/api.js
, the api.config.general.flatFileDirectory
directive is where your web site's "root" is. You can also use actions to manipulate file content with the api.staticFile.get
method. actionhero is also a great choice to power your front-end applications (angular.js, ember, etc). The examples below are purposefully sparse and often eschew convention and best practices in favor of legibility. No external JS (jQuery, etc) is required to use actionhero in your website (although they will make your life much easier).
Provided in index.html
is a simple page which demonstrates how simple it is to call an action from the web to document the API we have created. If you visit the the showDocumentation
action (generated with a new project), actionhero will describe it's capabilities, and we can then render them on our web page.
Sockets
relevant documentation section:
While this application probably makes the most sense being used in a web browser, actionhero can still provide a TCP/Socket API for clients who wish to use it. There is nothing new you need to do to enable it.
Let's add a comment with the socket API and Telnet. We'll make some mistakes so you can see how error responses are handled. You will also notice how with TCP clients, params are 'sticky' by default, and you only need to set them once per session.
> telnet localhost 5000
Trying 127.0.0.1...
Connected to localhost.
Escape character is '^]'.
{"welcome":"Hello! Welcome to the actionhero api","room":"defaultRoom","context":"api"}
paramAdd userName=evan
{"status":"OK","context":"response","data":null,"messageCount":1}
paramAdd title=first-post
{"status":"OK","context":"response","data":null,"messageCount":2}
paramsView
{"status":"OK","context":"response","data":{"limit":100,"offset":0,"userName":"evan","title":"first-post"},"messageCount":3}
postsList
{"posts":["first-post"],"context":"response","messageCount":4}
paramAdd comment=a-comment-from-tcp
{"status":"OK","context":"response","data":null,"messageCount":5}
commentAdd
{"error":"Error: commenterName is a required parameter for this action","context":"response","messageCount":6}
paramAdd commenterName=tcp-client
{"status":"OK","context":"response","data":null,"messageCount":7}
commentAdd
{"context":"response","messageCount":8}
commentsView
{"comments":[{"comment":"a-comment-from-tcp","createdAt":1372223802362,"commentId":"tcp-client1372223802362"}],"context":"response","messageCount":9}
exit
Websockets
files discussed in this section:
relevant documentation section:
/public/chat.js
demonstrates how to use actionhero's websockets. The websocket
is a first-class protocol in acitonHero and has all the capabilities of web
and socket
. Like socket
, it is a persistent connection which also enables actionhero's chat room features. We will make use of them here.
Note that we include both the javascript actionheroWebSocket.js
Note how we make use of the event libraries of actionheroWebsocket
and build our events around it:
A = new actionheroWebSocket;
A.events = {
say: function(message){
A.log(message);
appendMessage(message);
},
alert: function(message){
alert(message.message);
},
welcome: function(message){
A.log(message);
appendMessage(message);
}
}
A.connect(function(err, details){
if(err != null){
A.log(err);
}else{
A.log("CONNECTED");
}
});
Tasks
files discussed in this section:
relevant documentation section:
actionhero comes with a robust task system for delayed / recurring tasks. For our example, we are going to create a task which will log some stats to the command line every 30 seconds. You can do much more with actionhero's task system, including distributed tasks, recurring tasks, and more.
./node_modules/.bin/actionhero generateTask --name=stats
- note how we set the
task.frequency
to run every 30 seconds - to enable our server to run tasks, we need to configure 'workers' to run. You can enable workers in
/config/tasks.js
. - to re-scheudle a job like ours, you will also need to enable the schduler process in
/config/tasks.js
.
Custom Server
files discussed in this section:
relevant documentation section:
One of the powers of actionhero is that you can build your own servers and transports. Think of a serer as any mechanism which creates a client which may then preform an action. This might be anything from an interface to an Arduino to a rabbitMQ client.
We will add a server which connects to the public twitter firehouse for tweets matching "#nodejs". We will take these tweets, and broadcast them to the chartroom we created above.
We will also use the ntwittter
package to help us with connecting to twitter
npm install ntwitter
(and add it to yourpackage.json
)- You will need to register a new twitter application at dev.twitter.com to generate the required access keys.
We created a new section in config/servers/twitter.js
in the api.config.servers
section to hold our twitter credentials:
exports.default = {
servers: {
twitter: function(api){
return {
enabled: true,
hashtag: "sports",
// Be sure to fill these in with your own API keys
consumer_key: "XXX",
consumer_secret: "XXX",
access_token_key: "XXX",
access_token_secret: "XXX",
}
}
}
}
- Generate a new custom server
./node_modules/.bin/actionhero generateServer --name=twitter
notes:
- we didn't actually allow these clients to join the chatroom (
attributes.canChat = false
) because we don't want them to really appear as members of the chat room. - we are however using the server-side chat api (
api.chatRoom.socketRoomBroadcast
) to send a custom message to the folks in thetwitter
chat room about the tweet
Next Steps / TODO
- Use cookie-based authentication rather than requiring the password and userName to be sent with each request
- Migrate to another database
- Implement a UI for the API
- Tests should be more inclusive, and test failure cases
- Pagination for all the
*view
actions