io-ts

What is io-ts?

The io-ts npm package is a TypeScript library that allows for the definition of runtime types, and the automatic validation of runtime values against those types. It leverages TypeScript's type system to ensure that data structures conform to specified schemas, providing a bridge between the runtime data and compile-time types.

What are io-ts's main functionalities?

Runtime type validation

This feature allows you to define a type and then validate an object against that type at runtime. If the object matches the type, the 'Right' branch is executed; otherwise, the 'Left' branch indicates a validation error.

{"const t = require('io-ts');\nconst User = t.type({\n  name: t.string,\n  age: t.number\n});\nconst result = User.decode({ name: 'Alice', age: 25 });\nif (result._tag === 'Right') {\n  console.log('Valid!', result.right);\n} else {\n  console.log('Invalid!', result.left);\n}"}

Type composition

io-ts allows for the composition of types, enabling complex type definitions by combining simpler ones. This is useful for building up the shape of data structures from reusable type components.

{"const t = require('io-ts');\nconst Name = t.string;\nconst Age = t.number;\nconst User = t.type({ name: Name, age: Age });\nconst result = User.decode({ name: 'Bob', age: 'not-a-number' });\n// result will be an instance of Left since 'age' is not a number"}

Custom types

io-ts allows the creation of custom types with additional validation logic. In this example, a 'PositiveNumber' type is created that only accepts positive numbers.

{"const t = require('io-ts');\nconst PositiveNumber = t.brand(\n  t.number,\n  (n): n is t.Branded<number, { readonly PositiveNumber: unique symbol }> => n > 0,\n  'PositiveNumber'\n);\nconst result = PositiveNumber.decode(-5);\n// result will be an instance of Left since the number is not positive"}

Other packages similar to io-ts

ajv

Ajv is a JSON schema validator that provides runtime data validation using predefined JSON schemas. It is similar to io-ts in that it validates data structures at runtime, but it uses JSON schema as the basis for validation rather than TypeScript types.

joi

Joi is an object schema validation library that allows for the description and validation of JavaScript objects. It is similar to io-ts in providing runtime validation, but it uses a fluent API for schema definition and does not integrate with TypeScript types in the same way.

yup

Yup is a JavaScript schema builder for value parsing and validation. It defines a schema using a declarative API and validates objects against the schema. Like io-ts, it provides runtime validation, but it does not leverage TypeScript's type system for type definitions.

class-validator

Class-validator allows for validation of class instances based on decorators. It is similar to io-ts in that it provides runtime validation, but it is designed to work with classes and decorators, offering a different approach to defining validation rules.

README

Table of contents

• Installation
• The idea
• TypeScript integration
• TypeScript compatibility
• Error reporters
• Custom error messages
• Implemented types / combinators
• Recursive types
  • Mutually recursive types
• Branded types / Refinements
• Exact types
• Mixing required and optional props
• Custom types
• Generic Types
• Piping
• Community
• Tips and Tricks
  • Union of string literals

Installation

To install the stable version:

npm i io-ts fp-ts

Note: fp-ts is a peer dependency for io-ts

The idea

A value of type Type<A, O, I> (called "codec") is the runtime representation of the static type A.

A codec can:

• decode inputs of type I (through decode)
• encode outputs of type O (through encode)
• be used as a custom type guard (through is)

class Type<A, O, I> {
  constructor(
    /** a unique name for this codec */
    readonly name: string,

    /** a custom type guard */
    readonly is: (u: unknown) => u is A,

    /** succeeds if a value of type I can be decoded to a value of type A */
    readonly validate: (input: I, context: Context) => Either<Errors, A>,

    /** converts a value of type A to a value of type O */
    readonly encode: (a: A) => O
  ) {}

  /** a version of `validate` with a default context */
  decode(i: I): Either<Errors, A>
}

The Either type returned by decode is defined in fp-ts, a library containing implementations of common algebraic types in TypeScript.

The Either type represents a value of one of two possible types (a disjoint union). An instance of Either is either an instance of Left or Right:

type Either<E, A> =
  | {
      readonly _tag: 'Left'
      readonly left: E
    }
  | {
      readonly _tag: 'Right'
      readonly right: A
    }

Convention dictates that Left is used for failure and Right is used for success.

Example

A codec representing string can be defined as:

import * as t from 'io-ts'

const string = new t.Type<string, string, unknown>(
  'string',
  (input: unknown): input is string => typeof input === 'string',
  // `t.success` and `t.failure` are helpers used to build `Either` instances
  (input, context) => (typeof input === 'string' ? t.success(input) : t.failure(input, context)),
  // `A` and `O` are the same, so `encode` is just the identity function
  t.identity
)

and we can use it as follows:

import { isRight } from 'fp-ts/lib/Either'

isRight(string.decode('a string')) // true
isRight(string.decode(null)) // false

More generally the result of calling decode can be handled using fold along with pipe (which is similar to the pipeline operator)

import * as t from 'io-ts'
import { pipe } from 'fp-ts/lib/pipeable'
import { fold } from 'fp-ts/lib/Either'

// failure handler
const onLeft = (errors: t.Errors): string => `${errors.length} error(s) found`

// success handler
const onRight = (s: string) => `No errors: ${s}`

pipe(
  t.string.decode('a string'),
  fold(onLeft, onRight)
)
// => "No errors: a string"

pipe(
  t.string.decode(null),
  fold(onLeft, onRight)
)
// => "1 error(s) found"

We can combine these codecs through combinators to build composite types which represent entities like domain models, request payloads etc. in our applications:

import * as t from 'io-ts'

const User = t.type({
  userId: t.number,
  name: t.string
})

So this is equivalent to defining something like:

type User = {
  userId: number
  name: string
}

The advantage of using io-ts to define the runtime type is that we can validate the type at runtime, and we can also extract the corresponding static type, so we don’t have to define it twice.

TypeScript integration

Codecs can be inspected:

This library uses TypeScript extensively. Its API is defined in a way which automatically infers types for produced values

Note that the type annotation isn't needed, TypeScript infers the type automatically based on a schema (and comments are preserved).

Static types can be extracted from codecs using the TypeOf operator:

type User = t.TypeOf<typeof User>

// same as
type User = {
  userId: number
  name: string
}

TypeScript compatibility

The stable version is tested against TypeScript 3.5.2

io-ts version	required TypeScript version
2.x+	3.5.2+
1.6.x+	3.2.2+
1.5.3	3.0.1+
1.5.2-	2.7.2+

Note. This library is conceived, tested and is supposed to be consumed by TypeScript with the strict flag turned on.

Note. If you are running < typescript@3.0.1 you have to polyfill unknown.

You can use unknown-ts as a polyfill.

Error reporters

A reporter implements the following interface

interface Reporter<A> {
  report: (validation: Validation<any>) => A
}

This package exports a default PathReporter reporter

Example

import { PathReporter } from 'io-ts/lib/PathReporter'

const result = User.decode({ name: 'Giulio' })

console.log(PathReporter.report(result))
// => [ 'Invalid value undefined supplied to : { userId: number, name: string }/userId: number' ]

You can define your own reporter. Errors has the following type

interface ContextEntry {
  readonly key: string
  readonly type: Decoder<any, any>
}

interface Context extends ReadonlyArray<ContextEntry> {}

interface ValidationError {
  readonly value: unknown
  readonly context: Context
}

interface Errors extends Array<ValidationError> {}

Example

import { pipe } from 'fp-ts/lib/pipeable'
import { fold } from 'fp-ts/lib/Either'

const getPaths = <A>(v: t.Validation<A>): Array<string> => {
  return pipe(
    v,
    fold(errors => errors.map(error => error.context.map(({ key }) => key).join('.')), () => ['no errors'])
  )
}

console.log(getPaths(User.decode({}))) // => [ '.userId', '.name' ]

Custom error messages

You can set your own error message by providing a message argument to failure

Example

import { either } from 'fp-ts/lib/Either'

const NumberFromString = new t.Type<number, string, unknown>(
  'NumberFromString',
  t.number.is,
  (u, c) =>
    either.chain(t.string.validate(u, c), s => {
      const n = +s
      return isNaN(n) ? t.failure(u, c, 'cannot parse to a number') : t.success(n)
    }),
  String
)

console.log(PathReporter.report(NumberFromString.decode('a')))
// => ['cannot parse to a number']

You can also use the withMessage helper from io-ts-types

Implemented types / combinators

Type	TypeScript	codec / combinator
null	null	t.null or t.nullType
undefined	undefined	t.undefined
void	void	t.void or t.voidType
string	string	t.string
number	number	t.number
boolean	boolean	t.boolean
unknown	unknown	t.unknown
array of unknown	Array<unknown>	t.UnknownArray
array of type	Array<A>	t.array(A)
record of unknown	Record<string, unknown>	t.UnknownRecord
record of type	Record<K, A>	t.record(K, A)
function	Function	t.Function
literal	's'	t.literal('s')
partial	Partial<{ name: string }>	t.partial({ name: t.string })
readonly	Readonly<A>	t.readonly(A)
readonly array	ReadonlyArray<A>	t.readonlyArray(A)
type alias	type T = { name: A }	t.type({ name: A })
tuple	[ A, B ]	t.tuple([ A, B ])
union	A | B	t.union([ A, B ])
intersection	A & B	t.intersection([ A, B ])
keyof	keyof M	t.keyof(M) (only supports string keys)
recursive types		t.recursion(name, definition)
branded types / refinements	✘	t.brand(A, predicate, brand)
integer	✘	t.Int (built-in branded codec)
exact types	✘	t.exact(type)
strict	✘	t.strict({ name: A }) (an alias of t.exact(t.type({ name: A })))

Recursive types

Recursive types can't be inferred by TypeScript so you must provide the static type as a hint

interface Category {
  name: string
  categories: Array<Category>
}

const Category: t.Type<Category> = t.recursion('Category', () =>
  t.type({
    name: t.string,
    categories: t.array(Category)
  })
)

Mutually recursive types

interface Foo {
  type: 'Foo'
  b: Bar | undefined
}

interface Bar {
  type: 'Bar'
  a: Foo | undefined
}

const Foo: t.Type<Foo> = t.recursion('Foo', () =>
  t.interface({
    type: t.literal('Foo'),
    b: t.union([Bar, t.undefined])
  })
)

const Bar: t.Type<Bar> = t.recursion('Bar', () =>
  t.interface({
    type: t.literal('Bar'),
    a: t.union([Foo, t.undefined])
  })
)

Branded types / Refinements

You can brand / refine a codec (any codec) using the brand combinator

// a unique brand for positive numbers
interface PositiveBrand {
  readonly Positive: unique symbol // use `unique symbol` here to ensure uniqueness across modules / packages
}

const Positive = t.brand(
  t.number, // a codec representing the type to be refined
  (n): n is t.Branded<number, PositiveBrand> => n >= 0, // a custom type guard using the build-in helper `Branded`
  'Positive' // the name must match the readonly field in the brand
)

type Positive = t.TypeOf<typeof Positive>
/*
same as
type Positive = number & t.Brand<PositiveBrand>
*/

Branded codecs can be merged with t.intersection

// t.Int is a built-in branded codec
const PositiveInt = t.intersection([t.Int, Positive])

type PositiveInt = t.TypeOf<typeof PositiveInt>
/*
same as
type PositiveInt = number & t.Brand<t.IntBrand> & t.Brand<PositiveBrand>
*/

Exact types

You can make a codec exact (which means that additional properties are stripped) using the exact combinator

const ExactUser = t.exact(User)

User.decode({ userId: 1, name: 'Giulio', age: 45 }) // ok, result is right({ userId: 1, name: 'Giulio', age: 45 })
ExactUser.decode({ userId: 1, name: 'Giulio', age: 43 }) // ok but result is right({ userId: 1, name: 'Giulio' })

Mixing required and optional props

You can mix required and optional props using an intersection

const A = t.type({
  foo: t.string
})

const B = t.partial({
  bar: t.number
})

const C = t.intersection([A, B])

type C = t.TypeOf<typeof C>

// same as
type C = {
  foo: string
} & {
  bar?: number | undefined
}

You can apply partial to an already defined codec via its props field

const PartialUser = t.partial(User.props)

type PartialUser = t.TypeOf<typeof PartialUser>

// same as
type PartialUser = {
  name?: string
  age?: number
}

Custom types

You can define your own types. Let's see an example

import { either } from 'fp-ts/lib/Either'

// represents a Date from an ISO string
const DateFromString = new t.Type<Date, string, unknown>(
  'DateFromString',
  (u): u is Date => u instanceof Date,
  (u, c) =>
    either.chain(t.string.validate(u, c), s => {
      const d = new Date(s)
      return isNaN(d.getTime()) ? t.failure(u, c) : t.success(d)
    }),
  a => a.toISOString()
)

const s = new Date(1973, 10, 30).toISOString()

DateFromString.decode(s)
// right(new Date('1973-11-29T23:00:00.000Z'))

DateFromString.decode('foo')
// left(errors...)

Note that you can deserialize while validating.

Generic Types

Polymorphic codecs are represented using functions. For example, the following typescript:

interface ResponseBody<T> {
  result: T
  _links: Links
}
interface Links {
  previous: string
  next: string
}

Would be:

// t.Mixed = t.Type<any, any, unknown>
const ResponseBody = <C extends t.Mixed>(codec: C) =>
  t.interface({
    result: codec,
    _links: Links
  })

const Links = t.interface({
  previous: t.string,
  next: t.string
})

And used like:

const UserModel = t.type({
  name: t.string
})

functionThatRequiresRuntimeType(ResponseBody(t.array(UserModel)), ...params)

Piping

You can pipe two codecs if their type parameters do align

const NumberCodec = new t.Type<number, string, string>(
  'NumberCodec',
  t.number.is,
  (s, c) => {
    const n = parseFloat(s)
    return isNaN(n) ? t.failure(s, c) : t.success(n)
  },
  String
)

const NumberFromString = t.string.pipe(
  NumberCodec,
  'NumberFromString'
)

Community

• io-ts@1.x
  • io-ts-types - A collection of codecs and combinators for use with io-ts
  • io-ts-reporters - Error reporters for io-ts
  • geojson-iots - codecs for GeoJSON as defined in rfc7946 made with io-ts
  • graphql-to-io-ts - Generate typescript and cooresponding io-ts types from a graphql schema
  • io-ts-promise - Convenience library for using io-ts with promise-based APIs

Tips and Tricks

Union of string literals

Use keyof instead of union when defining a union of string literals

const Bad = t.union([
  t.literal('foo'),
  t.literal('bar'),
  t.literal('baz')
  // etc...
])

const Good = t.keyof({
  foo: null,
  bar: null,
  baz: null
  // etc...
})

Benefits

• unique check for free
• better performance, O(log(n)) vs O(n)

Beware that keyof is designed to work with objects containing string keys. If you intend to define a numbers enumeration, you have to use an union of number literals :

const HttpCode = t.union([
  t.literal(200),
  t.literal(201),
  t.literal(202)
  // etc...
])

Changelog

2.0.1

• Bug Fix
  • fix getTags algorithm for mutually recursive codecs, closes #354 (@gcanti)