fluent-json-schema

fluent-json-schema

A fluent API to generate JSON schemas (draft-07) for Node.js and browser. Framework agnostic.

Features

• Fluent schema implements JSON Schema draft-07 standards
• Faster and shorter way to write a JSON Schema via a fluent API
• Runtime errors for invalid options or keywords misuse
• JavaScript constants can be used in the JSON schema (e.g. enum, const, default ) avoiding discrepancies between model and schema
• TypeScript definitions
• Coverage 99%

Install

npm install fluent-json-schema --save

or

yarn add fluent-json-schema

Usage

const S = require('fluent-json-schema')

const ROLES = {
  ADMIN: 'ADMIN',
  USER: 'USER',
}

const schema = S.object()
  .id('http://foo/user')
  .title('My First Fluent JSON Schema')
  .description('A simple user')
  .prop('email', S.string().format(S.FORMATS.EMAIL).required())
  .prop('password', S.string().minLength(8).required())
  .prop('role', S.string().enum(Object.values(ROLES)).default(ROLES.USER))
  .prop(
    'birthday',
    S.raw({ type: 'string', format: 'date', formatMaximum: '2020-01-01' }) // formatMaximum is an AJV custom keywords
  )
  .definition(
    'address',
    S.object()
      .id('#address')
      .prop('line1', S.anyOf([S.string(), S.null()])) // JSON Schema nullable
      .prop('line2', S.string().raw({ nullable: true })) // Open API / Swagger  nullable
      .prop('country', S.string())
      .prop('city', S.string())
      .prop('zipcode', S.string())
      .required(['line1', 'country', 'city', 'zipcode'])
  )
  .prop('address', S.ref('#address'))

console.log(JSON.stringify(schema.valueOf(), undefined, 2))

Schema generated:

{
  "$schema": "http://json-schema.org/draft-07/schema#",
  "definitions": {
    "address": {
      "type": "object",
      "$id": "#address",
      "properties": {
        "line1": {
          "anyOf": [
            {
              "type": "string"
            },
            {
              "type": "null"
            }
          ]
        },
        "line2": {
          "type": "string",
          "nullable": true
        },
        "country": {
          "type": "string"
        },
        "city": {
          "type": "string"
        },
        "zipcode": {
          "type": "string"
        }
      },
      "required": ["line1", "country", "city", "zipcode"]
    }
  },
  "type": "object",
  "$id": "http://foo/user",
  "title": "My First Fluent JSON Schema",
  "description": "A simple user",
  "properties": {
    "email": {
      "type": "string",
      "format": "email"
    },
    "password": {
      "type": "string",
      "minLength": 8
    },
    "birthday": {
      "type": "string",
      "format": "date",
      "formatMaximum": "2020-01-01"
    },
    "role": {
      "type": "string",
      "enum": ["ADMIN", "USER"],
      "default": "USER"
    },
    "address": {
      "$ref": "#address"
    }
  },
  "required": ["email", "password"]
}

TypeScript

With "esModuleInterop": true activated in the tsconfig.json:

import S from 'fluent-json-schema'

const schema = S.object()
  .prop('foo', S.string())
  .prop('bar', S.number())
  .valueOf()

With "esModuleInterop": false in the tsconfig.json:

import * as S from 'fluent-json-schema'

const schema = S.object()
  .prop('foo', S.string())
  .prop('bar', S.number())
  .valueOf()

Validation

Fluent schema does not validate a JSON schema. However, many libraries can do that for you. Below a few examples using AJV:

npm install ajv --save

or

yarn add ajv

Validate an empty model

Snippet:

const ajv = new Ajv({ allErrors: true })
const validate = ajv.compile(schema.valueOf())
let user = {}
let valid = validate(user)
console.log({ valid }) //=> {valid: false}
console.log(validate.errors) //=> {valid: false}

Output:

{valid: false}
errors: [
  {
    keyword: 'required',
    dataPath: '',
    schemaPath: '#/required',
    params: { missingProperty: 'email' },
    message: "should have required property 'email'",
  },
  {
    keyword: 'required',
    dataPath: '',
    schemaPath: '#/required',
    params: { missingProperty: 'password' },
    message: "should have required property 'password'",
  },
]

Validate a partially filled model

Snippet:

user = { email: 'test', password: 'password' }
valid = validate(user)
console.log({ valid })
console.log(validate.errors)

Output:

{valid: false}
errors:
[ { keyword: 'format',
    dataPath: '.email',
    schemaPath: '#/properties/email/format',
    params: { format: 'email' },
    message: 'should match format "email"' } ]

Validate a model with a wrong format attribute

Snippet:

user = { email: '[e-mail]', password: 'password' }
valid = validate(user)
console.log({ valid })
console.log('errors:', validate.errors)

Output:

{valid: false}
errors: [ { keyword: 'required',
    dataPath: '.address',
    schemaPath: '#definitions/address/required',
    params: { missingProperty: 'country' },
    message: 'should have required property \'country\'' },
  { keyword: 'required',
    dataPath: '.address',
    schemaPath: '#definitions/address/required',
    params: { missingProperty: 'city' },
    message: 'should have required property \'city\'' },
  { keyword: 'required',
    dataPath: '.address',
    schemaPath: '#definitions/address/required',
    params: { missingProperty: 'zipcoce' },
    message: 'should have required property \'zipcode\'' } ]

Valid model

Snippet:

user = { email: '[e-mail]', password: 'password' }
valid = validate(user)
console.log({ valid })

Output:

{valid: true}

Extend schema

Normally inheritance with JSON Schema is achieved with allOf. However when .additionalProperties(false) is used the validator won't understand which properties come from the base schema. S.extend creates a schema merging the base into the new one so that the validator knows all the properties because it is evaluating only a single schema. For example, in a CRUD API POST /users could use the userBaseSchema rather than GET /users or PATCH /users use the userSchema which contains the id, createdAt and updatedAt generated server side.

const S = require('fluent-json-schema')
const userBaseSchema = S.object()
  .additionalProperties(false)
  .prop('username', S.string())
  .prop('password', S.string())

const userSchema = S.object()
  .prop('id', S.string().format('uuid'))
  .prop('createdAt', S.string().format('time'))
  .prop('updatedAt', S.string().format('time'))
  .extend(userBaseSchema)

console.log(userSchema)

Selecting certain properties of your schema

In addition to extending schemas, it is also possible to reduce them into smaller schemas. This comes in handy when you have a large Fluent Schema, and would like to re-use some of its properties.

Select only properties you want to keep.

const S = require('fluent-json-schema')
const userSchema = S.object()
  .prop('username', S.string())
  .prop('password', S.string())
  .prop('id', S.string().format('uuid'))
  .prop('createdAt', S.string().format('time'))
  .prop('updatedAt', S.string().format('time'))

const loginSchema = userSchema.only(['username', 'password'])

Or remove properties you dont want to keep.

const S = require('fluent-json-schema')
const personSchema = S.object()
  .prop('name', S.string())
  .prop('age', S.number())
  .prop('id', S.string().format('uuid'))
  .prop('createdAt', S.string().format('time'))
  .prop('updatedAt', S.string().format('time'))

const bodySchema = personSchema.without(['createdAt', 'updatedAt'])

Detect Fluent Schema objects

Every Fluent Schema object contains a boolean isFluentSchema. In this way, you can write your own utilities that understands the Fluent Schema API and improve the user experience of your tool.

const S = require('fluent-json-schema')
const schema = S.object().prop('foo', S.string()).prop('bar', S.number())
console.log(schema.isFluentSchema) // true

Documentation

• Full API Documentation.
• JSON schema reference.

Acknowledgments

Thanks to Matteo Collina for pushing me to implement this utility! 🙏

Related projects

• JSON Schema Draft 7
• Understanding JSON Schema (despite is referring to draft 6 the guide still good to grasp the main concepts)
• AJV JSON Schema validator
• jsonschema.net an online JSON Schema visual editor (it doesn't support advance features)

Licence

Licensed under MIT.