schema-nozzle generates JSON Schema from TypeScript interfaces, type aliases, classes, and enums. When developing a RESTful API server using frameworks like Express.js or Fastify.js, you might have experienced the need to write both JSON Schema and TypeScript types. Although the code is different, the meaning is the same, making it feel like doing the same work twice. To solve this problem, tools like json-schema-to-ts are used. Schema-nozzle does the reverse of json-schema-to-ts, automating the process of converting TypeScript types into JSON Schema.

Why schema-nozzle?

• Generates JSON Schema from TypeScript types.
• Integrates well with documentation tools like JSDoc and TypeDoc as it uses TypeScript types.
• When using Fastify, it can automate validation and automatically generate Swagger documentation.
• Supports various shapes of generating JSON Schema.

Strict JSON data validations are need many effort. You can reduce effort using schema-nozzle and Feel free 🤩!

• Getting Started
• Installation
• How it works?
  • TypeScript interface
  • json-schema
• Usage
• Performance
• Example using fastify.js
• Recommand option
• Relate To
• Roadmaps
• License

npx schema-nozzle init
npx schema-nozzle refresh

You can create configuration and list file using init command. And you can run refresh command, schema-nozzle generate json-schema from interface, type alias, class and enum.

You can see this mechanics!

npm install schema-nozzle --save-dev

schema-nozzle using TypeScript Compiler API. So schema-nozzle exactly know interface, type alias, class and enum.

graph LR

INP_TI[interface] --> SN[schema-nozzle]
INP_TT[type alias] --> SN[schema-nozzle]
INP_TC[class] --> SN[schema-nozzle]
INP_TE[enum] --> SN[schema-nozzle]
SN --> |exported| DB[db.json]
SN --> |not exported| IG[ignored]

• schema-nozzle generate json-schema using ts-json-schema-generator
• .nozzlefiles file follow gitignore spec.
• only generated exported interface, type alias, class and enum

Here is real example,

This is input source file.

export interface I18nDto {
  /** i18n resource id */
  id: string;

  /**
   * iso639-1 language code
   *
   * @minLength 2
   * @maxLength 5
   * */
  language: string;

  /** i18n resource content */
  content: string;

  /**
   * i18n resource use on
   *
   * @minItems 1
   * @maxItems 10
   * */
  used?: string[];
}

This is output json-schema.

{
  "I18nDto": {
    "id": "I18nDto",
    "filePath": "I18nDto.ts",
    "dependency": {
      "import": {
        "name": "I18nDto",
        "from": []
      },
      "export": {
        "name": "I18nDto",
        "to": [
          "IForeginStudentDto33",
          "IForeginStudentDto",
          "IReqReadStudentQuerystring",
          "IStudentDto",
          "IProfessorDto"
        ]
      }
    },
    "schema": {
      "$id": "I18nDto",
      "$schema": "http://json-schema.org/draft-07/schema#",
      "type": "object",
      "properties": {
        "id": {
          "type": "string",
          "description": "i18n resource id"
        },
        "language": {
          "type": "string",
          "description": "iso639-1 language code",
          "minLength": 2,
          "maxLength": 5
        },
        "content": {
          "type": "string",
          "description": "i18n resource content"
        },
        "used": {
          "type": "array",
          "items": {
            "type": "string"
          },
          "description": "i18n resource use on",
          "minItems": 1,
          "maxItems": 10
        }
      },
      "required": ["id", "language", "content"],
      "additionalProperties": false
    }
  }
}

You can use this schema like that,

// use ajv schema store
import Ajv from 'ajv';

const ajv = new Ajv();
const db = JSON.parse((await fs.readFile('db.json')).toString());
Object.values(db).forEach((item) => ajv.addSchema(item.schema));

const validator = ajv.compile({ $ref: 'I18nDto' });

or

// don't use schema store
import Ajv from 'ajv';

const ajv = new Ajv();
const db = JSON.parse((await fs.readFile('db.json')).toString());
const validator = ajv.compile(db['I18nDto'].schema);

You can see help from --help option.

# display help for each commands
npx schema-nozzle --help

# display help for add commands
npx schema-nozzle add --help

# display help for del commands
npx schema-nozzle del --help

# display help for refresh commands
npx schema-nozzle refresh --help

# display help for truncate commands
npx schema-nozzle truncate --help

# display help for watch commands
npx schema-nozzle watch --help

Also you can see detail option here.

0.19.0 version enhance performance. Now remove pain point from mass schema generation.

This image is result of 388 schema extraction using M1 macbook pro(16GB RAM, 1TB SSD, 10core). schema-nozzle spent only 6.14 second! 🙌 🙆

A complete example of using schema-nozzle to create a swagger.io document and use json-schema to process input-output value verification can be found at Ma-eum. See the example of how DTO type declaration handles swagger.io document creation, json-schema creation, and typedoc document creation all at once!

• fastify with schema-nozzle example

extraTags and additionalProperties options enable in .nozzlerc

{
  "tsconfig": "./tsconfig.json",
  "list-file": "./.nozzlefiles",
  "cli-logo": true,
  "generatorOption": {
    "additionalProperties": true,
    "extraTags": ["example"]
  }
}

additionalProperties option permit additional properties in request object. And extraTags option can add to example field for swagger.io document.

• ts-json-schema-generator
  • json-schema generator
• ts-morph
  • TypeScript Compiler API wrapper

• add watch command: watch .nozzlefiles list and add/del schema
• enhance init command: find varity name of tsconfig. eg. tsconfig.*.json
• tag support each schema
• load, get, set interface for schema store
• documentation site
• $id field enhance: enclude directory path like #/greeting/hello/world
• add more test

This software is licensed under the MIT.