Inspired by restful-react

Summary

• Code Generation
  • Usage
  • Samples
  • Validation of the OpenAPI specification
  • Import from GitHub
  • Transforming an Original Spec
  • Advanced configuration
    • Config File Format
    • Config File Example

Code Generation

orval is able to generate client with appropriate type-signatures (TypeScript) from any valid OpenAPI v3 or Swagger v2 specification, either in yaml or json formats.

Usage

Type-safe data fetchers can be generated from an OpenAPI specification using the following command:

• orval --input MY_OPENAPI_SPEC.yaml --output my-awesome-generated-types.tsx

This command can be invoked by either:

• Installing orval globally and running it in the terminal: npm i -g orval, or
• Adding a script to your package.json like so:

      "scripts": {
        "start": "webpack-dev-server",
        "build": "webpack -p",
+       "generate-fetcher": "orval --input MY_SWAGGER_DOCS.json --output FETCHERS.tsx"
      }

Your client can then be generated by running npm run generate-fetcher. Optionally, we recommend linting/prettifying the output for readability like so:

      "scripts": {
        "start": "webpack-dev-server",
        "build": "webpack -p",
        "generate-fetcher": "orval --input MY_SWAGGER_DOCS.json --output FETCHERS.tsx",
+       "postgenerate-fetcher": "prettier FETCHERS.d.tsx --write"
      }

Samples

You can find below some samples on codesandbox

• react app
• react app with react query
• angular app

Validation of the OpenAPI specification

To enforce the best quality as possible of specification, we have integrated the amazing OpenAPI linter from IBM. We strongly encourage you to setup your custom rules with a .validaterc file, you can find all useful information about this configuration here.

Import from GitHub

Using an url in input like this orval --input https://github.com/OAI/OpenAPI-Specification/blob/master/examples/v3.0/petstore.yaml allows us to create your client from an OpenAPI spec remotely hosted on GitHub. ^{(how is this real life 🔥 )}

To generate components from remote specifications, you'll need to follow the following steps:

1. Visit your GitHub settings.

2. Click Generate New Token and choose the following:

   Token Description: (enter anything)
   Scopes:
       [X] repo
           [X] repo:status
           [X] repo_deployment
           [X] public_repo
           [X] repo:invite
   

3. Click Generate token.

4. Copy the generated string.

5. Open a terminal and run orval --input https://github.com/OAI/OpenAPI-Specification/blob/master/examples/v3.0/petstore.yaml, substituting things where necessary.

6. You will be prompted for a token.

7. Paste your token.

8. You will be asked if you'd like to save it for later. This is entirely up to you and completely safe: it is saved in your node_modules folder and not committed to version control or sent to us or anything: the source code of this whole thing is public so you're safe.

   Caveat: Since your token is stored in node_modules, your token will be removed on each npm install of orval.

9. You're done! 🎉

Transforming an Original Spec

In some cases, you might need to augment an existing OpenAPI specification on the fly, for code-generation purposes.

Transfomer example:

// /path/to/my-transformer.js

/**
 * Transformer function for orval.
 *
 * @param {OpenAPIObject} schema
 * @return {OpenAPIObject}
 */
module.exports = (inputSchema) => ({
  ...inputSchema,
  // Place your augmentations here
  paths: Object.entries(schema.paths).reduce(
    (mem, [path, pathItem]) => ({
      ...mem,
      [path]: Object.entries(pathItem).reduce(
        (pathItemMem, [verb, operation]) => ({
          ...pathItemMem,
          [verb]: {
            ...fixOperationId(path, verb, operation),
          },
        }),
        {},
      ),
    }),
    {},
  ),
});

Advanced configuration

orval supports the concept of "schema stitching" in a RESTful ecosystem as well. We are able to tie multiple backends together and generate code using a single configuration file, orval.config.js

To activate this "advanced mode", replace all flags from your orval call with the config flag: --config orval.config.js (or any filename that you want).

⚠️ Note: using a config file makes use of all of the options contained therein, and ignores all other CLI flags.

Config File Format

interface RestfulClientConfig {
  [backend: string]: {
    // path or output options object
    output?: string | OutputOptions;
    // path, url, or input options object
    input?: string | InputOptions;
  };
}

type OverrideInput = {
  transformer?: string;
};

interface InputOptions = {
    // path or url to the openapi spec
    target?: string;
    // validation of your openapi spec
    validation?: boolean;
    // override the input that's give you the possibility to add whatever you want to your openapi spec
    override?: OverrideInput;
};

type OutputClient = 'axios' | 'angular' | 'react-query';

export type OutputMode = 'single' | 'split' | 'tags' | 'tags-split';

type MockProperties =
  | { [key: string]: unknown }
  | ((specs: OpenAPIObject) => { [key: string]: unknown });

type OperationOptions = {
  transformer?: string;
  mutator?: string;
  mock?: {
    data?: MockProperties;
    properties?: MockProperties;
  };
};


type OverrideOutput = {
  transformer?: string;
  mutator?: string;
  operations?: { [key: string]: OperationOptions };
  mock?: {
    properties?: MockProperties;
  };
};

interface OutputOptions = {
  // override title
  title?: (title: string) => string
  // default axios
  client?: OutputClient
  // path to the file which will contains the implementation
  target?: string;
  // path to the directory that will contains your models (if not define the target will contains the schemas)
  schemas?: string;
  // single to have everything in the same file
  // split to have definition, implementation, schemas, mock in differents files
  // tags to have a file by tag
  // tags-split a mix of tags and split
  mode?: OutputMode;
  // add mock to your implementation
  // old version will be removed in the next major version
  mock?: boolean | 'old-version';
  // override the output like your mock implementation or transform the api implementation like you want
  override?: OverrideOutput;
};

Config File Example

// orval.config.js
module.exports = {
  'petstore-file': {
    input: 'examples/petstore.yaml',
    output: 'examples/petstoreFromFileSpecWithConfig.ts',
  },
  'petstore-file-transfomer': {
    output: {
      target: 'examples/petstoreFromFileSpecWithTransformer.ts',
      schemas: 'examples/model',
      mode: 'split',
      mock: true,
    },
    input: {
      target: 'examples/petstore.yaml',
      transformer: 'examples/transformer-add-version.js',
    },
    override: {
      // contains operationId of your spec with override options
      operations: {
        listPets: {
          // mutator the output of your api call
          mutator: 'examples/transformer-response-type.js',
          mock: {
            // override mock properties
            properties: () => {
              return {
                id: faker.random.number({ min: 1, max: 9 }),
              };
            },
          },
        },
        showPetById: {
          mock: {
            // override mock for this api call
            data: () => ({
              id: faker.random.number({ min: 1, max: 99 }),
              name: faker.name.firstName(),
              tag: faker.helpers.randomize([faker.random.word(), undefined]),
            }),
          },
        },
      },
      mock: {
        // override mock properties for all api calls
        properties: {
          '/tag|name/': 'jon',
        },
      },
    },
  },
};

// package.json
{
  "scripts": {
    "gen": "orval --config orval.config.js",
    "gen-first": "orval --config orval.config.js myFirstBackend"
  }
}