A fluent API to generate JSON schemas (draft-07) for Node.js and browser. Framework agnostic.
- 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%
npm i fluent-json-schema
or
yarn add fluent-json-schema
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"]
}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()A named export is also available to work with native ESM modules:
import { S } from 'fluent-json-schema'
const schema = S.object()
.prop('foo', S.string())
.prop('bar', S.number())
.valueOf()Fluent schema does not validate a JSON schema. However, many libraries can do that for you. Below a few examples using AJV:
npm i ajv
or
yarn add ajv
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'",
},
]
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"' } ]
Snippet:
user = { email: '[email protected]', 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\'' } ]
Snippet:
user = { email: '[email protected]', password: 'password' }
valid = validate(user)
console.log({ valid })Output:
{valid: true}
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)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'])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) // trueThanks to Matteo Collina for pushing me to implement this utility! 🙏
- 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)
Licensed under MIT.
