📘️ openapi-typescript
View examples:
Usage
CLI
🗄️ Reading specs from file system
npx openapi-typescript schema.yaml --output schema.ts
# 🤞 Loading spec from tests/v2/specs/stripe.yaml…
# 🚀 schema.yaml -> schema.ts [250ms]
☁️ Reading specs from remote resource
npx openapi-typescript https://petstore.swagger.io/v2/swagger.json --output petstore.ts
# 🤞 Loading spec from https://petstore.swagger.io/v2/swagger.json…
# 🚀 https://petstore.swagger.io/v2/swagger.json -> petstore.ts [650ms]
Thanks to @psmyrdek for the remote spec feature!
Using in TypeScript
Import any top-level item from the generated spec to use it. It works best if you also alias types to save on typing:
import { components } from './generated-schema.ts';
type APIResponse = components["schemas"]["APIResponse"];
The reason for all the ["…"]
everywhere is because OpenAPI lets you use more characters than are valid TypeScript identifiers. The goal of this project is to generate all of your schema, not merely the parts that are “TypeScript-safe.”
Also note that there’s a special operations
interface that you can import OperationObjects
by their operationId:
import { operations } from './generated-schema.ts';
type getUsersById = operations["getUsersById"];
This is the only place where our generation differs from your schema as-written, but it’s done so as a convenience and shouldn’t cause any issues (you can still use deep references as-needed).
Thanks to @gr2m for the operations feature!
stdout
Outputting to npx openapi-typescript schema.yaml
Generating multiple schemas
In your package.json
, for each schema you’d like to transform add one generate:specs:[name]
npm-script. Then combine
them all into one generate:specs
script, like so:
"scripts": {
"generate:specs": "npm run generate:specs:one && npm run generate:specs:two && npm run generate:specs:three",
"generate:specs:one": "npx openapi-typescript one.yaml -o one.ts",
"generate:specs:two": "npx openapi-typescript two.yaml -o two.ts",
"generate:specs:three": "npx openapi-typescript three.yaml -o three.ts"
}
If you use npm-run-all, you can shorten this:
"scripts": {
"generate:specs": "run-p generate:specs:*",
You can even specify unique options per-spec, if needed. To generate them all together, run:
npm run generate:specs
Rinse and repeat for more specs.
For anything more complicated, or for generating specs dynamically, you can also use the Node API.
CLI Options
Option | Alias | Default | Description |
---|---|---|---|
--output [location] |
-o |
(stdout) | Where should the output file be saved? |
--prettier-config [location] |
(optional) Path to your custom Prettier configuration for output | ||
--raw-schema |
false |
Generate TS types from partial schema (e.g. having components.schema at the top level) |
Node
npm i --save-dev openapi-typescript
const { readFileSync } = require("fs");
const swaggerToTS = require("openapi-typescript").default;
const input = JSON.parse(readFileSync("spec.json", "utf8")); // Input can be any JS object (OpenAPI format)
const output = swaggerToTS(input); // Outputs TypeScript defs as a string (to be parsed, or written to a file)
The Node API is a bit more flexible: it will only take a JS object as input (OpenAPI format), and return a string of TS definitions. This lets you pull from any source (a Swagger server, local files, etc.), and similarly lets you parse, post-process, and save the output anywhere.
If your specs are in YAML, you’ll have to convert them to JS objects using a library such as js-yaml. If you’re batching large folders of specs, glob may also come in handy.
Migrating from v1 to v2
Project Goals
- Support converting any OpenAPI 3.0 or 2.0 (Swagger) schema to TypeScript types, no matter how complicated
- The generated TypeScript types must match your schema as closely as possible (i.e. don’t convert names to
PascalCase
or follow any TypeScript-isms; faithfully reproduce your schema as closely as possible, capitalization and all) - This library is a TypeScript generator, not a schema validator.
Contributing
PRs are welcome! Please see our CONTRIBUTING.md guide. Opening an issue beforehand to discuss is encouraged but not required.
✨
Contributors Thanks goes to these wonderful people (emoji key):
This project follows the all-contributors specification. Contributions of any kind welcome!