TODO BEFORE MERGE:

• [ ] ability to generate code using programmatic API (handy for converting old schemas)
• [ ] allow objects to redefine keywords
• [ ] dates (Joi.date)
• [ ] support refs
• [ ] resolve refs when generating
• [ ] custom keywords
• [ ] docs!

Typed-JSON-Schema

TypeScript-Friendly JSON-Schema Definitions

This library is able to validate JSON-Schema at runtime, and also emit type definitions for the validated data. This means that if you define your data using our JSON-schema builder, you automatically get TypeScript safety on those types.

Examples

Schema generation:

await validator.validate(string.minLength(3), 'asdf') // type: string

await validator.validate(array(string.pattern(/\w+/))) // result type: string[]

await validator.validate(object.required('a', 'b').properties({
  a: string,
  b: array(number) 
}) // result type: {a: string, b: number[] }

Validation

import { schema, string, Validator } from 'typed-json-schema'

const validator = new Validator()

const StringOrNull = schema.type(['string', 'null'])

const result = await validator.validate(StringOrNull, 'Hello'),

result.toFixed(1) // error: Property 'toFixed' does not exist on type 'string'.
result.toLowerCase() //error: object is possibly null
result && validation.result.toLowerCase() //success!

Usage

Schema Creation

In general, all JSON-Schema keywords can be used as builders on the schema object.

For example:

import { schema } from 'typed-json-schema'

const mySchema = schema
  .type('string')
  .minLength(3)
  .pattern(/regex/)

Refer to JSON-Schema keywords for a list of available keywords.

Shortcuts:

import { string, boolean, number, array, object }

const mySchema = string //same as schema.type('string')
const mySchema = number //same as schema.type('number')
const mySchema = boolean //same as schema.type('boolean')
const mySchema = array(string) //same as schema.type('array').items(string)
const mySchema = object({ a: string, b: array(number) }) //same as schema.type('object').properties({ a: string, b: array(number) })

Validation:

Use a Validator to validate a json schema:

import { schema, Validator }

const validator = new Validator({ coerceTypes: true }) //any AJV options can be supplied
const result = await validator.validate(schema.type('number')) //result is a number

see AJV for a list of options that can be passed to the validator.