The Koa Web Server

In this worksheet you will be learning the basic principles involved in building a multipage website using the Koa web server and the JavaScript language. In previous worksheets you have been running a koa server but in this lab we will be taking a deep dive into its inner workings and will also be learning how to program using the JavaScript language.

For this lab Assume this is being covered as the first step in learning JS.

1 JavaScript

The third skill you will need, after HTML5 and CSS3 is the ability to control the operation of your website by writing scripts using a suitable programming language. There are many languages to choose from such as Python, PHP or Java however in this module you will be using the JavaScript language.

There are a huge range of language features but this worksheet will focus on those that are directly relevent to the assignment. You are encouraged to study the language in more depth. It is assumed you are comfortable with core programming principles using the Python language.

Javascript (more correctly known as ECMAscript) has been around since the mid 1990s and has acquired a bad press due to some of its badly implemented features that can introduce bugs into your code. In the last couple of years it has changed considerably and version 6 often referred to as ECMA6 (the version you will be using) is now considered a flexible, powerful scripting language.

You will sometimes hear different names used for this language:

JavaScript is the original name when the language was developed by Netscape.
ECMAScript is the name of the language standard developed by ECMA, from the original Javascript implementation.
NodeJS is JavaScript running on a web server using the language intepreter from the Chrome Web Browser.

Basically these are all implementations of the same language but in this chapter we will focus on programming using NodeJS since it offers a more consistent interpreter and you will find it the most useful for your websites. In later chapters we will move JavaScript to the web browser.

Warning: most online tutorials have been written using the older flavours of JavaScript. You should not rely on these giving you the correct information.

We must avoid using this older style and this worksheet will only cover the most up to date implementation of the language.

1.1 NodeJS

Traditionally JavaScript ran in the web browser and was used to carry out tasks such as animation and form validation. Recently however the Chrome JavaScript interpreter (called V8) has been modified to run on the server which means we can use JavaScript as a replacement for the older server-side languages such as PHP.

There are a number of benefits over other web scripting languages:

The first, and most obvious answer, is that, as a web developer, you are going to have to learn JavaScript anyway because its the only language that will run natively in a web browser so why not use the same language for both client and server?

The second benefit, and one we will return to in later chapters, is that, unlike other scripting languages which create a new process for each connected user, NodeJS runs a single process shared between all users. Since processes are expensive in computing resources it means that a NodeJS deployment is far more scalable. In chapter 3 we will learn how it handles concurrency through the use of threads.

The downside of only having a single process is that the script can only handle thing at a time. To write efficient scripts it is therefore important to avoid waiting for jobs to complete and the language makes use of concepts such as asynchronous callbacks to improve efficiency. We will be covering this important topic later in the worksheet.

1.2 Packages

One of the most valuable features of NodeJS is its package manager which allows you to install additional functionality in the form of packages. There are thousands of these to choose from, fully documented on the Node Package Manager website. Later on you will be shown how you can publish your own packages.

Packages can be installed either locally or globally.

Local packages are installed in a node_modules/ directory within the directory containing the NodeJS scripts. This is the way we install most of the modules. These are only available within that directory.
Global packages on the other hand are installed system-wide and can be accessed by all the scripts. Normally these need to be installed using root privileges. We will be using global packages in a later chapter. We won't be installing many scripts in this way.

We install packages using the npm install command. Modules can be deleted using npm uninstall and you can see a list of the modules you have installed in the current directory using npm ls.

Start by opening the index.js script in the

Use the terminal tool to access the exercises/06_routing/ directory and ope

TODO: Koa example of installing uninstalling and listing packages.

2 Koa Web Server

We will be using the Koa framework as our web server in this module. Koa was designed by the team who created the Express Framework which was one of the first mainstream web servers to use the JavaSscript language. Almost all the online tutorials cover the use of Express so why have we chosen not to use this?

The Express framework has grown over the years and contains a lot of functionaity that we won't need. Koa has the opposite philosophy. It’s a minimalist framework where all the functionality is aplit into smaller modules that can be imported when needed. . Koa is Express 5.0 in spirit; it’s by the same people that created Express, and the authors changed the name to avoid upsetting people because of the backwards incompatible changes (which also happened in the Express 3 to Express 4 transition).

The only issue is that there was a big change from Koa v1 to Koa v2 and the older tutorials are therefore of no use. We will focus on using the latest version of Koa, currently 2.7 at the time of writing. Lets take an in-depth look at a script which should be familiar to you to understand how the web server works.

2.1 The Manifest

A correctly-configured NodeJS application includes a special manifest file called package.json. You will find one in the exercises/04_koa/ directory. This contains important metadata about the application. If you open this you will notice it contains a JSON string which includes a number of keys:

The name key contains the name of the application, in this case koa-intro.
Underneath this you will find the version number and a brief description of the software.
The main key defined the entry point to the application (the script that should be run).
Next are script aliases, shortcuts to run different commands. In our example we have a test alias that is used to trigger the jest testing tool and a server alias to start the server.
After the scripts key you will notice a jest key that contains the configuration options for the Jest testing tools.
Next we have the author and licence fields.
Finally the manifest list the packages needed for the application to run:
1. The dependencies key contains all the packages needed to run the application such as koa and its plugins.
2. The devDependencies key lists all the extra packages needed when developing the application but not needed by the app itself such as the testing tools.

To see how the manifest can help us with running an application:

Navigate to the exercises/04_koa/ directory using terminal.
Use the npm install --production command to install all the packages needed to run the application (no development tools).
Use the npm run server to trigger the server alias.
Try using npm run test to run the test suite, notice that the jest command is no currently installed.
Run npm install to install the developer tools (it won't' install the production modules because these are already installed).
Now you can run the test suite.

2.2 The Routing File

Study the index.js script in the exercises/04_koa/ directory.

The first line is the shebang, it tells the script what application is needed to run it. If the file is executable you can simply call this without using the node command like this: ./index.js instead of node index.js.
Lines 3-6 are where we import all the modules we are going to be using.
Lines 8-12 are where we create instances of the koa and router objects and configure the plugins (known as the middleware).
Next we define any global variables and constants. In this script we store the port number as a constant.
The main part of the script defines the routes and we will be covering these in more detail as we progress through the lab.
Right at the end we start the server on the defined port and export the koa object app. By exporting it we can import the script into our automated test suite (briefly covered in the previous lab).

2.2.1 Test Your Understanding

Lets start with a quick refresher:

Create a new route called /test and create a file called mytest.html containing a valid HTML5 web page that displays your name in the views/ directory. Restart the server and make sure the page displays.
Create a new CSS3 stylesheet called test.css in the css/ directory in the public/ directory. Link it to your new page and change the font face and colour of the text.
Find a download a photo of the University, add it to the images/ directory inside the public/ directory and make sure you can view this in the browser.

Contents:

Package Manifests
Routes
Request and response objects
Templating
Modular code

1 Package Manifests

A package manifest is an json-formatted file called package.json that is created in the root directory of a NodeJS application. I describes the application (name, author, etc), the entry point (the script that needs to be run to launch the application) and identifies the third-party packages (including their specific versions) needed to run the application. It also supports development be defining the packages needed to support the code development and also allows us to define command _aliases) which means we don't need to type in complex commands.

The real benefit to having dependencies defined like this in package.json, is that it becomes possible to install the correct versions of all the required packages with a single command. This means that we could use an automated build and deploy tool.

You have probably spotted another file called package-lock.json. This contains a list of all packages installed indicating their dependencies plus details of all dependencies.

1.1 Understanding the Manifest

Let's look at a simple example. You can find this in the examples/06_express/todo/ directory. Open the package.json file:

The first few keys define the project name, author name, etc.
The main key contains the script that should be run to start the application, in this example, index.js is the entry point to our application. This is mainly used by automated build-deply tools so they know how to start the app.
Next there is a scripts object. This is where we can define script aliases. These store complex commands and allow them to be executed using shorter commands:
1. try running npm run hello.
2. Can you see what has happened?
3. Try npm run start.
4. Because this is such a useful command we can abbreviate it to npm start, try this.
After this there is a dependencies object. This lists all the packages needed for the application to run.
1. Instead of installing each package separately try the command npm install --only=production.
2. Now lets see what packages were installed using npm list --depth=0. You should see that the express package was installed. The package version should match that specified in the dependencies object.
The final object is called devDependencies and contains all the packages needed to develop the application.
1. Lets install these packages using npm install --only=dev. This will install the eslint package.
2. List the locally-installed modules again using npm list --depth=0 to make sure the package was installed.

2 Routing

Start by locating the exercises/06_express/todo/ directory and locate the index.js file. This is the routing file used by the express web server.
The first few lines import the package and configure the server:
1. Lines 3-4 import the express package and create an instance called app.
2. Next the port number is stored in a constant. You should always handle numbers by assigning to constants to make their purpose clear.
Next, on lines 9-11 we define a route. This has two parameters:
1. The path to match. In this case, the / represents the base url with no additional segments.
2. The function to run if this route is accessed. In this case it loads the contents of the coventry.html file and sends it back to the web browser.
Finally we tell express to listen on the specified port. There are two parameters:
1. The port.
2. A function to run as soon as the server is ready to receive requests.
3. Now open the SSH Terminal and navigate to the directory, install the express package and run the index.js file.
4. Finally you need to open a browser tab and navigate to the base URL on the correct port. You will see the following:

2.1 Routes

Every request sent from the client is handled by a route. The server compares the requested HTTP method and route against the strings passed as the first parameter until it finds a match. If there are no routes that match the specific URL the koa server will repond with a 404 NOT FOUND response.

When a match is found, the server runs the callback (anonymous function) that has been supplied as the second parameter. This function takes a single ctx parameter which is an object that contains all the data from both the request and response (the context).

app.get('/test', ctx => {
    // code goes here
})

2.2 The Request Object

The ctx parameter contains a request object (ctx.request) that contains all the data passed as part of the HTTP request headers and body. it contains all the information from these headers, in particular, given the request:

http://www.example.com/hello/mark?gender=male

Object	Contains	Example
`ctx.request.query`	The querystring	`gender=male`
`ctx.request.body`	The request body	-
`ctx.request.hostname`	The server hostname	`www`
`ctx.request.baseUrl`	The base URL	`www.example.com`
`ctx.request.path`	The route	`/hello/mark`
`ctx.request.ip`	The server IP address	`192.168.0.1`
`ctx.request.params`	The parameters	`/mark`

ctx.request.accepts(types) Checks if the specified content types are acceptable, based on the request’s Accept HTTP header field. The method returns the best match, or if none of the specified content types is acceptable, returns false (in which case, the application should respond with 406 "Not Acceptable"). req.accepts('html')

ctx.request.get(field) Returns the specified HTTP request header field (case-insensitive match). The Referrer and Referer fields are interchangeable. req.get('Content-Type')

2.3 The Response Object

The response object contains the data to be returned to the client as the HTTP response. It contains a number of functions.

Function	Description	Example
`res.write()`	adds text to the response body	`res.write('hello world')`
`res.send()`	sends text to the response body and sends to client	`res.send('hello world')`
`res.setHeader()`	adds a new response header	`res.setHeader('content-type', 'text/html')`
`res.sendFile()`	sends the contents of a file to the client	`res.sendFile(\`${__dirname}/form.html`)`
`res.status()`	sets the HTTP status	`res.status(201)`
`res.end()`	sends the current response body to the client	`res.end()`

currency/ example

3 Modular Code

Un until now, all your JavaScript/NodeJS code has been in a single file (commonly named index.js). Whilst this works for small scripts, as your scripts get longer it becomes increasingly difficult to manage. The solution is to modularise your code by splitting it up into several different files.

3.1 Reducing the Size of the Routes File

Your first step should be to remove as much code as possible from your main routes file. The file should handle the http requests and send back the response but nothing else. let's look at an example.

Open the books/index.js file. This file contains 2 routes, /bad and /good. Start the server and view the /bad route which should display a form which allows you to search for a book. Try searching for books on different topics to see how this works.

As you search:

Examine the URL used and identify any query parameters.
Study the index.html template file to understand how the data is rendered.
Study the code in the routes file. Notice there are two nested callbacks.

You have probably noticed that there is a lot of code in the /bad route! This code does 2 different tasks:

It contains the business logic to make API requests and tidy the data.
It also takes this data and sends it back to the browser as an HTTP response.

This is why the code is hard to read. It therefore makes the code difficult to maintain and new features are likely to introduce bugs. Later on, when we start automating our code testing we would quickly discover this code is very difficult to write tests for. To fix this we need to separate out the business logic from the routing.

3.2 The Exports Object

Every NodeJS file includes a special module object that represents the current module. It contains a nested object called exports. Anything in this exports object will be exposed (public) whenever this module is imported and used.

// basic.js

module.exports.name = 'John Doe'

module.exports.hello = (name, callback) => {
  // code goes here.
  return callback(null, `hello ${name}`)
}

In this example, the name property is available to any script importing this module. We also have a function expression in the hello property.

To use this we need to import this basic.js module.

const basic = import('./basic')

console.log(basic.hello)

basic.hello('Mark', (err, data) => {
  if(err) console.log('an error has occurred')
  console.log(data)
})

Let's look at a more useful example! Open the books/books.js file. Notice that there is a function expression stored in the module.exports.searchByString property. This means it is visible outside the module. This takes the data from the request and returns books based on the search query, it does not add this data to the template or return it to the client.

Now look at the index.js file and locate the /good route. You can see immediately that there is a lot less code. The code here calls the searchByString() function expression and then uses the callback to add the data to the template and send the response to the client.

3.3 Private Functions

Any function in a module that has not been added to the module.exports object is visible only to other code in the module (private scope). locate the buildString function. Notice that this is a standard function that is used to build the correct URL. It is used in the betterSearchString function literal and replaces a block of inline code.

3.4 Test Your Understanding

Modify the template to display the data as a 3 column table with the title in the first column, the ISBN in the second and the unique id in the third.
Add a hyperlink to the title to link to the route /book/XXXXX, where XXXXX is the unique ID of that book.
1. Now remove the third column.
2. Clicking on these links will display a message 'page not found'.
Create a new route to handle this, the route will be /book/:id. This should initially just display the id of the book on the page.
1. You can access the ID value in your script using the req.params.id object.
Create and export a new function literal in the books.js file, this should retrieve the book title, description and author(s) and display these:
1. Use the URL https://www.googleapis.com/books/v1/volumes/XXXXX to retrieve the book details (where XXXXX is the book ID).
2. Print this to the console to identify the structure.
3. Extract the data and print to the console.
4. Build a new html template called book.html to display the information.
Create a back button to return to the search results.
1. How can you ensure the search results are still there when you click on this button?
Can you replace the callbacks with promises/async functions?
Can you split the code logic into multiple functions?

TEACHING-MATERIALS/04 Intro to Koa.md