Skip to content

Getting Started with TypeSpec For REST APIs

Introduction

Welcome to our tutorial on using TypeSpec to define REST APIs with HTTP. In this section, we’ll introduce you to TypeSpec, help you set up your environment, and cover the basic syntax and structure of TypeSpec. By the end of this section, you’ll have a solid foundation to build upon in the subsequent sections.

What is TypeSpec?

TypeSpec is a language and toolset developed by Microsoft for defining data models and service APIs. It provides a structured way to describe the shape and behavior of data and services, ensuring consistency and reducing errors in API development. With TypeSpec, you can generate code, documentation, and other artifacts from your API definitions, making it easier to maintain and evolve your services. Microsoft uses TypeSpec internally to define APIs for various products and services, including Azure.

TypeSpec is used to define the interface of your API, which clients will use to interact with resources provided by your service. This includes specifying the operations, request and response models, and error handling mechanisms. The actual API logic is implemented in the backend service, which processes the requests and communicates with the database.

Before we start writing TypeSpec code, we need to set up our development environment. For detailed instructions on setting up your environment, please refer to the Installation Guide.

Summary of Setup and Installation

  1. Install Node.js: Download and install Node.js from nodejs.org. This will also install npm, the Node.js package manager. The minimum versions required are Node.js 20.0.0 and npm 7.0.0.
  2. Install TypeSpec CLI: Run npm install -g @typespec/compiler to install the TypeSpec CLI.
  3. Verify Installation: Run tsp --version to verify that the TypeSpec CLI is installed correctly.
  4. Create a New Project:
    • Run tsp init and select the Generic REST API template.
    • Run tsp install to install dependencies.
    • Run tsp compile . to compile the initial file.
    • Run tsp compile . --watch to automatically compile changes on save.

Project Structure Overview

Once you’ve completed these steps, you’ll have a basic TypeSpec project set up. Here’s an overview of the files and directories in your TypeSpec project:

  • main.tsp
  • tspconfig.yaml
  • package.json
  • Directorynode_modules/
    • …
  • Directorytsp-output/
    • Directory@typespec/
      • Directoryopenapi3/
        • openapi.yaml
  • main.tsp: Entry point for TypeSpec definitions.
  • tspconfig.yaml: TypeSpec compiler configuration.
  • package.json: Project metadata and dependencies.
  • node_modules/: Installed dependencies.
  • tsp-output/: Generated files.
  • openapi.yaml: Generated OpenAPI specification.

As we work through the tutorial, keep the openapi.yaml file open in Visual Studio or VS Code to watch the API specification evolve as we make changes.

Basic Syntax and Structure

Now that we have our environment set up, let’s dive into the basic syntax and structure of TypeSpec. We’ll create a simple REST API for a pet store by introducing concepts in a layered fashion, increasing complexity as we progress through the tutorial.

As the tutorial advances and the code examples grow more complex, we’ll highlight changes in the code to help you easily spot where new lines have been added.

Import and Using Statements

Before defining models and services, we need to import the necessary TypeSpec libraries and make them available in our namespace.

As we progress through the tutorial, you can follow along by updating the main.tsp file in your project and compiling the changes to see the results reflected in the generated openapi.yaml specification.

You can also alternatively use the Try it feature with the code samples to quickly view the generated OpenAPI spec in your browser via the TypeSpec Playground.

Let’s begin by adding the following import and using statements to the main.tsp file:

Try it
main.tsp
import "@typespec/http";


using TypeSpec.Http;

In this example:

  • import statement brings in the TypeSpec HTTP library, which provides the decorators and models we’ll be using to define our REST API.
  • using statement makes the imported library available in the current namespace, allowing us to use its features and decorators.

NOTE: Your generated project file likely already has these import/using statements, plus import/using for the @typespec/openapi3 library. The @typespec/openapi3 library is necessary for emitting the OpenAPI specification file but is not required for creating our Pet Store API in TypeSpec. Remove them from your main.tsp file so your code matches the example above.

Defining a REST Service

A REST service in TypeSpec is defined using the @service decorator. This decorator allows you to specify metadata about your service, such as its title. Additionally, you can use the @server decorator to define the server endpoint where your service will be hosted.

Example: Defining a Service with a Title and Server Endpoint

Let’s start by defining a simple REST service for a Pet Store:

Try it
main.tsp
import "@typespec/http";


using TypeSpec.Http;
// highlight-start
@service({
  title: "Pet Store",
})
@server("https://example.com", "Single server endpoint")
// highlight-end

In this example:

  • The @service decorator is used to define a service with the title “Pet Store”.
  • The @server decorator specifies the server endpoint for the service, which is “https://example.com”.

OpenAPI Comparison: In OpenAPI, this is similar to defining the info object (which includes the title) and the servers array (which includes the server URL).

NOTE: This code will not compile as-is because we’ve not yet defined a namespace for these decorators to apply to. We’ll cover that topic next.

Organizing with Namespaces

Namespaces in TypeSpec help you organize your models and operations logically. They act as containers for related definitions, making your API easier to manage and understand.

Example: Creating a Namespace

Let’s create a namespace for our Pet Store service:

Try it
main.tsp
import "@typespec/http";


using TypeSpec.Http;


@service({
  title: "Pet Store",
})
@server("https://example.com", "Single server endpoint")


// highlight-next-line
namespace PetStore;

In this example:

  • The namespace keyword is used to define a top-level namespace named PetStore.
  • All models and operations related to the Pet Store service will be defined within this namespace.
  • The first use of namespace defines the top-level namespace and does not require brackets. This is because it serves as the primary container for all related definitions.
  • Any subsequent namespaces defined within this top-level namespace will require brackets to indicate that they are nested within the top-level namespace.

OpenAPI Comparison: In OpenAPI, namespaces are similar to using tags to group related operations and definitions.

Defining Models

In TypeSpec, a model is a fundamental building block used to define the structure of data. Models are used to represent entities, such as a Pet, with various properties that describe the entity’s attributes.

Example: Defining a Simple Model

Let’s define a simple model for a Pet:

Try it
main.tsp
import "@typespec/http";


using TypeSpec.Http;


@service({
  title: "Pet Store",
})
@server("https://example.com", "Single server endpoint")
namespace PetStore;


// highlight-start
model Pet {
  id: int32;
  name: string;
  age: int32;
  kind: petType;
}


enum petType {
  dog: "dog",
  cat: "cat",
  fish: "fish",
  bird: "bird",
  reptile: "reptile",
}
// highlight-end

In this example:

  • The model keyword is used to define a new model named Pet.
  • The Pet model has four properties: id, name, age, and kind.
  • The petType enum defines possible values for the kind property.

OpenAPI Comparison: In OpenAPI, this is similar to defining a schema object under the components section, where you define the structure and properties of your data models.

Example: Adding Validation Annotations

We can add validation annotations to our model properties to enforce certain constraints:

Try it
main.tsp
import "@typespec/http";


using TypeSpec.Http;


@service({
  title: "Pet Store",
})
@server("https://example.com", "Single server endpoint")
namespace PetStore;


model Pet {
  id: int32;


  // highlight-next-line
  @minLength(1)
  name: string;


  // highlight-next-line
  @minValue(0)
  // highlight-next-line
  @maxValue(100)
  age: int32;


  kind: petType;
}


enum petType {
  dog: "dog",
  cat: "cat",
  fish: "fish",
  bird: "bird",
  reptile: "reptile",
}

In this example:

  • @minLength(1) ensures that the name property has at least one character.
  • @minValue(0) and @maxValue(100) ensure that the age property is between 0 and 100.

OpenAPI Comparison: In OpenAPI, this is similar to using minLength, minimum, and maximum constraints within the schema object.

Conclusion

In this section, we introduced you to TypeSpec, set up the development environment, and covered basic language syntax and structure. We defined a simple REST service, organized our API using namespaces, and defined a model with validation annotations.

With this foundational knowledge, you’re now ready to dive deeper into defining operations and handling different types of responses in your REST API. In the next section, we’ll expand our API by adding CRUD operations.