Web APIs

An API, or application programming interface, is a set of rules that allow one application to communicate with another application.

Web APIs are how applications communicate with other applications over a network. Throughout the remainder of this chapter, we will explore web APIs and a pattern for organizing them called REST.

REST is an application design pattern, not unlike MVC.

REST, however, relies on an abstract concept called application state that we’ll cover in more detail on the following pages. Here, we dive into what differentiates all web APIs from MVC applications.

MVC Without the V

Web APIs are actually similar to MVC web applications but for one major distinction: Web APIs are not concerned with the presentation of data. A web API encompasses the model and controller aspects of MVC, but is not responsible for the view layer. In a web API, the view, or presentation of data, is decoupled from the model and controller that manage and transfer data. This separation leads to the development of two different applications, a client (front-end) and a web API (back-end).

The separation between client and web API provides the following benefits:

• Client applications can be developed to operate on a range of platforms (web, mobile, CLI, GUI, etc).
• Client and web API applications can be developed by different, specialized programming teams.
• Client and web API applications can be hosted on separate infrastructure.

Client Interacts With Data

Despite being separated from each other, the client application still relies on interactions with application data through the web API. A front-end client must ultimately present data to an end-user. This means the client must request a representation of the data from the web API. After the client receives the representation, it is parsed, styled and rendered to the user.

Responsibilities of a Web API

The chief responsibility of a web API is to exchange representations of data with a corresponding client application. APIs are also gatekeepers, maintaining rules about how end-users may interact with application data.

Data Delivery

Think about how a view works in MVC.

1. Data is injected into a template file,
2. That template is rendered in HTML in the browser when a controller action method is invoked,
3. The data is sent back to the user

We call this approach server-side rendering. A client application and web API work in a similar way. However, instead of injecting the data into a template on the server, data is transferred over the network through AJAX requests made by the client.

When a client application receives data, it injects that data into its HTML using JavaScript. This all occurs from within a browser. This approach is called client-side rendering because the web API only sends data, the HTML is assembled on the user’s end.

Management of Data

A web API manages data by modeling objects that line up with the underlying business data. This is actually the same process we saw in MVC. Our models are class files that drive the interactions in our codebase. Web APIs often take advantage of ORMs just like MVC applications do.

If you have underlying classes that map to a database, you can easily make data available for use within a web API codebase.

Transference of Data

Beyond managing data, a web API also handles transferring data. A client application will make a request for some data. Our web API must contain controller files that can handle the requests. As a part of handling the request, the controller file must understand the request, access the requested data, package the data in an accepted format, and send the package as a response to the client application.

Here’s an overview of the steps to transfer data between a web API and client application:

1. A request for data comes from a client application to the API.
2. A controller file in the API catches the request.
3. The controller determines if the request is valid.
4. The controller transfers data from the database to an object, via the ORM.
5. The controller transforms the object into a package the client application can work with.
6. The controller responds to the client with the packaged data.

Representation of Data

Presentation vs Representation

As mentioned above, the client application presents the data to the end-user. However, the client relies on consuming a representation of data from the web API. Presentation is the rendered combination of data and visual styling intended for end-users.

The client application needs to know what format the data is in so that it can be transformed into a human readable presentation (HTML/CSS). A web API packages data into a format the client application accepts.

This format is called the representation of the data. The client application team and the web API team must agree to the underlying data format. A best practice is to use a universal representation widely accepted by client applications.

Universal Representation

It is necessary to adopt a universal representation because web APIs and client applications may be written in two different programming languages. Your web API may be written in C#/ASP.NET but the client application may be written using JavaScript and React.

While there are many languages and frameworks available in web development, they all support the creation and parsing of JSON. JSON is a standard in web development because it is simple to process in any language, compatible with HTTP, and seamlessly represents the structure of data objects.

JSON

JSON is currently a universal representation of data accepted by client applications. This means our web API packages data requested by a client application as JSON. The web API also transfers this JSON in its communication with a client application.

Let’s revisit the last two steps from our web API work flow

1. The controller transforms the object into a JSON representation.
2. The controller responds to the client with the JSON representation.

Let’s look at how exactly a client application makes a request and how a web API responds.

HTTP as The Language of Web APIs

HTTP is the protocol used for communication between a web API and a client application. Web APIs communicate over a network. The most common protocol of the internet is HTTP, so it comes as no surprise that HTTP is the language of web APIs. Similarly, our MVC applications use HTTP as the protocol for an end-user to access the application. HTTP also facilitates the communication between a client application and a web API.

Here’s a refresher on the basics of HTTP :

• Is a stateless request/response protocol.
• Requests and responses may include HTTP bodies.
• Responses always contain a three digit HTTP status code.
• Requests and responses always include HTTP headers.

We call HTTP a stateless protocol. State can be a complex concept that refers to a number of things. We’ll explore some aspects of it in more depth on the next page.

In the context of HTTP, think of state as information about application data that is transferred via HTTP bodies, HTTP status codes, and HTTP headers.

Bodies

An HTTP body can contain a large number of different media types, known as MIME types . A MIME type is associated with the HTTP header Content-Type. This header instructs the recipient of the HTTP request/response on what MIME type the HTTP body contains. We’ve seen a Content-Type:text/html HTTP header before.

 1
 2
 3
 4
 5
 6
 7
 8
 9
10

<!DOCTYPE html>
<html>
   <head>
      <title>My Web Page</title>
      content
   </head>
   <body>
      content
   </body>
</html>

1
2
3
4
5
6
7
8

{
   "title": "An Astronaut's Guide to Life on Earth",
   "author": "Chris Hadfield",
   "ISBN": 9780316253017,
   "year_published": 2013,
   "subject": ["Hadfield, Chris", "Astronauts", "Biography"],
   "available": true
}