swagger-client

What is swagger-client?

The swagger-client npm package is a JavaScript client for Swagger/OpenAPI specifications. It allows you to interact with APIs described by Swagger/OpenAPI documents, making it easier to consume and interact with RESTful services.

What are swagger-client's main functionalities?

Initialize Client

This feature allows you to initialize a Swagger client by providing the URL to the Swagger/OpenAPI specification. The client can then be used to interact with the API described by the specification.

const Swagger = require('swagger-client');
const client = new Swagger({ url: 'http://petstore.swagger.io/v2/swagger.json' });

Make API Calls

This feature allows you to make API calls using the initialized client. In this example, the client is used to call the 'getPetById' operation from the Petstore API.

client.then(client => {
  client.apis.pet.getPetById({ petId: 1 }).then(response => {
    console.log(response);
  });
});

Handle Authentication

This feature allows you to handle authentication by providing authorization details when initializing the client. In this example, an API key is provided for authentication.

const client = new Swagger({
  url: 'http://petstore.swagger.io/v2/swagger.json',
  authorizations: {
    api_key: new Swagger.ApiKeyAuthorization('api_key', 'YOUR_API_KEY', 'query')
  }
});

Other packages similar to swagger-client

axios

Axios is a promise-based HTTP client for the browser and Node.js. While it does not specifically target Swagger/OpenAPI specifications, it can be used to make HTTP requests to any API. Compared to swagger-client, axios is more general-purpose and does not provide built-in support for Swagger/OpenAPI documents.

openapi-client-axios

openapi-client-axios is a package that combines the capabilities of Axios with OpenAPI specifications. It allows you to generate an Axios client from an OpenAPI document, similar to swagger-client. However, it leverages Axios for making HTTP requests, providing a more flexible and widely-used HTTP client.

swagger-js

swagger-js is another JavaScript client for Swagger/OpenAPI specifications. It provides similar functionality to swagger-client, allowing you to interact with APIs described by Swagger/OpenAPI documents. The main difference is in the implementation and API design, but both packages serve the same purpose.

README

Swagger Client

Swagger Client is a JavaScript module that allows you to fetch, resolve, and interact with Swagger/OpenAPI documents.

New!

This is the new version of swagger-js, 3.x. The new version supports Swagger 2.0 as well as OpenAPI 3.

Want to learn more? Check out our FAQ.

For the older version of swagger-js, refer to the 2.x branch.

Note:

The npm package is called swagger-client and the GitHub repository is swagger-js. We'll be consolidating that soon. Just giving you the heads up. You may see references to both names.

Usage

Prerequisites

• Runtime:
  • browser: es5 compatible. IE11+
  • node v4.x.x
• Building
  • node v6.x.x

Download via npm

npm install swagger-client

Import in code

import Swagger from 'swagger-client'
// Or commonjs
const Swagger = require('swagger-client')

Import in browser

<script src='//unpkg.com/swagger-client' type='text/javascript'></script>
<script>
var swaggerClient = new SwaggerClient(specUrl);
</script>

API

This lib exposes these functionalities for Swagger 2.0 and OpenAPI 3:

• Static functions for...
  • HTTP Client
  • Document Resolver (monolithic & subtree)
  • TryItOut Executor
• A constructor with the methods...
  • HTTP Client, for convenience
  • Document Resolver, which will use url or spec from the instance
  • TryItOut Executor, bound to the http and spec instance properties
  • Tags Interface, also bound to the instance

HTTP Client

Swagger.http(req) exposes a Fetch-like interface with a twist: allowing url in the request object so that it can be passed around and mutated. It extends Fetch to support request and response interceptors and performs response & header serialization. This method could be overridden to change how SwaggerJS performs HTTP requests.

// Fetch-like, but support `url`, `query` and `xxxInterceptor`
const request = {
  url,
  query,
  method,
  body,
  headers,
  requestInterceptor,
  responseInterceptor,
  userFetch
}

Swagger.http(request)
  .then((res) => {
    res.statusCode // status code
    res.statusText // status text, ie: "Not Found"
    res.body       // JSON object or undefined
    res.obj        // same as above, legacy
    res.text       // textual body, or Blob
    res.headers    // header hash
  })
  .catch((err) => {
    err            // instanceof Error
    err.response   // response or null
  })

// Interceptors
Swagger.http({
  requestInterceptor: (req: Request) => Request | Promise<Request>
  responseInterceptor: (res: Response) => Response | Promise<Response>
})

// Custom Fetch
Swagger.http({
  userFetch: (url: String, options: Object) => Promise
})

Swagger Specification Resolver

Swagger.resolve({url, spec, http}) resolves $refs (JSON-Refs) with the objects they point to.

Swagger.resolve({url, spec, http}).then((resolved) => {
  resolved.errors // resolution errors, if any
  resolved.spec   // the resolved spec
})

This is done automatically if you use the constructor/methods

TryItOut Executor

An HTTP client for OAS operations, maps an operation and values into an HTTP request.

const params = {
  spec,

  operationId, // Either operationId, or you can use pathName + method
  (pathName),
  (method),

  parameters, // _named_ parameters in an object, eg: { petId: 'abc' }
  securities, // _named_ securities, will only be added to the request, if the spec indicates it. eg: {apiKey: 'abc'}
  requestContentType,
  responseContentType,

  (http), // You can also override the HTTP client completely
  (userFetch), // Alternatively you can just override the fetch method (if you want to use request.js or some other HttpAgent)
}

// Creates a request object compatible with HTTP client interface.
// If `pathName` and `method`, then those are used instead of operationId. This is useful if you're using this dynamically, as `pathName` + `method` are guarenteed to be unique.
const res = Swagger.execute({...params})

// You can also generate just the request ( without executing... )
const req = Swagger.buildRequest({...params})

Constructor and methods

Resolve the spec and expose some methods that use the resolved spec:

• Swagger(url, opts): Promise
• Exposes tags interface (see above)
• Exposes the static functions: execute, http, resolve and some other minor ones
• Exposes #http, #execute and #resolve bound to the instance

Swagger('http://petstore.swagger.io/v2/swagger.json')
  .then( client => {
      client.spec // The resolved spec
      client.originalSpec // In case you need it
      client.errors // Any resolver errors

      // Tags interface
      client.apis.pet.addPet({id: 1, name: "bobby"}).then(...)

      // TryItOut Executor, with the `spec` already provided
      client.execute({operationId: 'addPet', parameters: {id: 1, name: "bobby") }).then(...)
   })

Tags Interface

A client for operations. We're currently using the apis[tag][operationId]:ExecuteFunction interface, which can be disabled entirely using Swagger({disableInterfaces: true}) if you don't need it.

OperationId's are meant to be unique within spec, if they're not we do the following:

• If a tag is absent, we use default as the internal tag
• If an operationId is missing, we deduce it from the http method and path, i.e. ${method}${path}, with non-alphanumeric characters escaped to _. See these tests (1, 2) for examples.
• If an operationId is duplicated across all operationIds of the spec, we rename all of them with numbers after the ID to keep them unique. You should not rely on this, as the renaming is non-deterministic. See this test for an example.

Swagger({ url: "http://petstore.swagger.io/v2/swagger.json" }).then((client) => {
    client
      .apis
      .pet // tag name == `pet`
      .addPet({ // operationId == `addPet`
        id: 1,
        body: {
          name: "bobby",
          status: "available"
        }
      })
      .then(...)
})

If you'd like to use the operationId formatting logic from Swagger-Client 2.x, set the v2OperationIdCompatibilityMode option:

Swagger({
  url: "http://petstore.swagger.io/v2/swagger.json",
  v2OperationIdCompatibilityMode: true
}).then((client) => {
  // do things as usual
})

OpenAPI 3.0

OpenAPI 3.0 definitions work in a similar way with the tags interface, but you may need to provide additional data in an options object for server variables and request bodies, since these items are not actual parameters:

Swagger({...}).then((client) => {
    client
      .apis
      .pet // tag name == `pet`
      .addPet({ // operationId == `addPet`
        id: 1
      }, {
        requestBody: {
          name: "bobby",
          status: "available"
        },
        server: "http://petstore.swagger.io/{apiPrefix}/", // this should exactly match a URL in your `servers`
        serverVariables: {
          apiPrefix: "v2"
        }
      })
      .then(...)
})

In Browser

If you need activate CORS requests, just enable it by withCredentials property at http

<html>
<head>
<script src='//unpkg.com/swagger-client' type='text/javascript'></script> 
<script>

var specUrl = 'http://petstore.swagger.io/v2/swagger.json'; // data urls are OK too 'data:application/json;base64,abc...'
SwaggerClient.http.withCredentials = true; // this activates CORS, if necessary

var swaggerClient = new SwaggerClient(specUrl)
      .then(function (swaggerClient) {
          return swaggerClient.apis.pet.addPet({id: 1, name: "bobby"}); // chaining promises
      }, function (reason) {
         console.error("failed to load the spec" + reason);
      })
      .then(function(addPetResult) {
         console.log(addPetResult.obj);
         // you may return more promises, if necessary
      }, function (reason) {
          console.error("failed on API call " + reason);
      });
</script>
</head>
<body>
  check console in browser's dev. tools
</body>
</html>

Compatibility

SwaggerJS has some legacy signature shapes.

Execute

Response shape

// swagger-js
{
  url,
  method,
  status,
  statusText,
  headers,

  data, // The textual content
  obj   // The body object
}

// New shape
{
  url,
  method,
  status,
  statusText,
  headers, // See note below regarding headers

  text,    // The textual content
  body,    // The body object
}

Serializing Headers

By default the instance version of #http serializes the body and headers. However, headers pose an issue when there are multiple headers with the same name. As such we've left the static version of http to not perform any serialization.

Build

npm install
npm run test         # run test
npm run test:unit:watch   # run test with change watching
npm run test:lint         # run lint
npm run build        # package to release
npm run build:umd:watch    # package with non-minified dist/index.js (for debugging)
npm run build:bundle # build browser version available at .../browser/index.js

Migration from 2.x

There has been a complete overhaul of the codebase. For notes about how to migrate coming from 2.x, please see Migration from 2.x

Security contact

Please disclose any security-related issues or vulnerabilities by emailing [e-mail], instead of using the public issue tracker.

Graveyard

For features known to be missing from 3.x please see the Graveyard