Highly composable decoding and validation library

栏目: IT技术 · 发布时间: 5年前

内容简介:A lightweight decoding and validation library.This library provides a set of composable functions for typesafe schema validation and data decoding. Because typescript is staticaly typed this can be particularly useful for message validation in, for example

TS Prove

A lightweight decoding and validation library.

About

This library provides a set of composable functions for typesafe schema validation and data decoding. Because typescript is staticaly typed this can be particularly useful for message validation in, for example, a HTTP route handler by ensuring that the correct data is available.

Install

npm install ts-prove
yarn add ts-prove

Documentation

Example

In the example below, we construct a proof for the type Person<{ name: { first: string, last: string }, age: number }> . This proof is a callback which accepts some unknown data and returns either Failure<[string, unknown]> or Success<[null, Person]>

import P, { ProofType } from 'ts-prove'

const proveName = P.shape({
  first: P.string,
  last: P.optional(P.string),
})

const provePerson = P.shape({
  name: proveName,
  age: P.number,
})

type Person = ProofType<typeof provePerson>
// { name: { first: string, last?: string }, age: number }

provePerson({ name: { first: 'Bob' }, age: 'Incorect string value' }) // [string, unknown]
provePerson({ name: { first: 'Bob' }, age: 10 }) // [null, Person]

This could then be used to validate a payload.

import { isProved } from 'ts-prove'

export const createPerson = (req) => {
  const data = person(req.body)
  if (!isProved(data)) return req.send({ status: 500, body: data[0] })
  return db.createPerson(data[1])
}

db.createPerson = (p: Person) => void

Validation

Every proof is curried and accepts either a function of type (x: unknown) => true | string or some other value. When supplied with a function, a proof returns an instance of itself narrowed by the constraints of the callback. In the example bellow we use this feature to further narrow the number type to numbers betwen 10 and 19.

const teenage = P.number((x) => x > 10 || 'To young')((x) => x < 19 || 'To old')

teenage(9) // ['To young', unknown]
teenage(20) // ['To old', unknown]
teenage(13) // [null, number]

// This could then be used in a structured proof
const teenager = P.shape({ name: nameProof, age: teenage })

Proofs

All the proofs provided by this library are constructed using the function P<ValueType>(): Proof<ValueType> . We can use P to construct are own proofs by providing a return type as the type argument.

P.string = P<string>((x) => typeof x === 'string' || 'Expected string')

Here is a generic proof which does a deep equal comparison using lodash.isequal function.

import P from 'ts-prove'
import isEqual from 'lodash.isequal'

const equalProof = <T extends any>(type: T) => P<T>((input) => isEqual(type, input) | `Not equal`)

API

Type definitions

Type Definition
Sucess<T> [ null , T ]
Failure [ string , unknown ]
Check< T = unknown > (value: T ) => true | string
Proof< R > <T *extends Check<R> | R* >(x: T ): T extends Check<R> ? Proof<R> : Success<R> | Failure

P: Proof

Accepts a return type argument and a validation function that returns a failure of type string or boolean true.

  • string: Proof< string >
  • number: Proof< number >
  • boolean: Proof< boolean >
  • symbol: Proof< symbol >
  • null: Proof< null >
  • undefined: Proof< undefined >

Two utility proofs that always match but that resolve to different types.

  • any: Proof< any >
  • unknown: Proof< unknown >
  • array:< T extends Proof<any> >(proof: T ): Proof <*ProofType< T > []* >

    • Accepts a proof T as an argument and returns a proof for arrays containing type T.

    • P.array(P.string)
      // string[]
  • shape:< T extends { [x: string]: Proof<any> } >(proof: T ): Proof < { [Key in keyof T ]: ProofType< T[Key] > } >

    • Accepts a key value object where every value is a proof and returns a proof for objects of that shape.

    • P.shape({ name: P.string, age: P.number })
      // { name: string, age: number }
  • optional:< T extends Proof<any> >(proof: T ): Proof <*ProofType< T > | undefined* >

    • For use in shapes when you want the return type to indicate and optional value. Optional accepts a proof as an argument and returns union with undefined. For any other cases use or .

    • P.shape({ name: P.string, age: P.optional(P.number) })
      // { name: string, age?: number }
  • or: < T extends Proof<any>[] >(...proofs: T ): Proof < ProofType< T[number] > >

    • Accepts any number of proofs as arguments and returns a union of thoses proofs if any one of them match a given value.

    • P.or(P.string, P.number, P.symbol)
      // string | number | symbol

Helper functions

  • isProved: < T extends any >( x : Failure | Success< T >): x is Success< T >

    • Type guard for the return type of a proof
  • check:( proof : Proof<any>): Check< unknown >

    • Used to convert a proof to a check

Type guard library coming soon...

This project follows the all-contributors specification. Contributions of any kind are welcome!


以上就是本文的全部内容,希望本文的内容对大家的学习或者工作能带来一定的帮助,也希望大家多多支持 码农网

查看所有标签

猜你喜欢:

本站部分资源来源于网络,本站转载出于传递更多信息之目的,版权归原作者或者来源机构所有,如转载稿涉及版权问题,请联系我们

密码朋克

密码朋克

[澳] 朱利安·阿桑奇 / Gavroche / 中信出版社 / 2017-10 / 42.00元

互联网已经在世界各地掀起了革命,然而全面的打击也正在展开。随着整个社会向互联网迁移,大规模监控计划也正在向全球部署。我们的文明已经来到一个十字路口。道路的一边通往一个承诺“弱者要隐私,强 者要透明”的未来,而另一边则通往一个极权的互联网,在那里,全人类的权力被转移给不受问责的间谍机构综合体及其跨国公司盟友。 密码朋克是一群倡导大规模使用强密码术以保护我们的基本自由免遭攻击的活动家。维基解密的......一起来看看 《密码朋克》 这本书的介绍吧!

图片转BASE64编码
图片转BASE64编码

在线图片转Base64编码工具

Base64 编码/解码
Base64 编码/解码

Base64 编码/解码

URL 编码/解码
URL 编码/解码

URL 编码/解码