Project Icon

zodios

TypeScript API开发框架 支持自动完成和类型验证

Zodios是一个TypeScript API开发框架,集成了axios和zod库。该框架提供简洁的API声明方式,支持IDE自动完成和类型安全特性。Zodios具备参数和响应模式验证功能,并包含插件系统,支持fetch适配器和身份验证自动注入。框架为React和Solid提供了@tanstack/query封装,同时保留了axios和express的核心功能。Zodios旨在简化API开发流程,提升开发效率和代码质量。

Zodios

Zodios logo

Zodios is a typescript api client and an optional api server with auto-completion features backed by axios and zod and express
Documentation

langue typescript npm GitHub GitHub Workflow Status

Bundle Size Bundle Size Bundle Size Bundle Size Bundle Size Bundle Size Bundle Size

https://user-images.githubusercontent.com/633115/185851987-554f5686-cb78-4096-8ff5-c8d61b645608.mp4

What is it ?

It's an axios compatible API client and an optional expressJS compatible API server with the following features:

  • really simple centralized API declaration

  • typescript autocompletion in your favorite IDE for URL and parameters

  • typescript response types

  • parameters and responses schema thanks to zod

  • response schema validation

  • powerfull plugins like fetch adapter or auth automatic injection

  • all axios features available

  • @tanstack/query wrappers for react and solid (vue, svelte, etc, soon)

  • all expressJS features available (middlewares, etc.)

Table of contents:

Install

Client and api definitions :

> npm install @zodios/core

or

> yarn add @zodios/core

Server :

> npm install @zodios/core @zodios/express

or

> yarn add @zodios/core @zodios/express

How to use it on client side ?

For an almost complete example on how to use zodios and how to split your APIs declarations, take a look at dev.to example.

Declare your API with zodios

Here is an example of API declaration with Zodios.

import { Zodios } from "@zodios/core";
import { z } from "zod";

const apiClient = new Zodios(
  "https://jsonplaceholder.typicode.com",
  // API definition
  [
    {
      method: "get",
      path: "/users/:id", // auto detect :id and ask for it in apiClient get params
      alias: "getUser", // optional alias to call this endpoint with it
      description: "Get a user",
      response: z.object({
        id: z.number(),
        name: z.string(),
      }),
    },
  ],
);

Calling this API is now easy and has builtin autocomplete features :

//   typed                     auto-complete path   auto-complete params
//     ▼                               ▼                   ▼
const user = await apiClient.get("/users/:id", { params: { id: 7 } });
console.log(user);

It should output

{ id: 7, name: 'Kurtis Weissnat' }

You can also use aliases :

//   typed                     alias   auto-complete params
//     ▼                        ▼                ▼
const user = await apiClient.getUser({ params: { id: 7 } });
console.log(user);

API definition format

type ZodiosEndpointDescriptions = Array<{
  method: 'get'|'post'|'put'|'patch'|'delete';
  path: string; // example: /posts/:postId/comments/:commentId
  alias?: string; // example: getPostComments
  immutable?: boolean; // flag a post request as immutable to allow it to be cached with react-query
  description?: string;
  requestFormat?: 'json'|'form-data'|'form-url'|'binary'|'text'; // default to json if not set
  parameters?: Array<{
    name: string;
    description?: string;
    type: 'Path'|'Query'|'Body'|'Header';
    schema: ZodSchema; // you can use zod `transform` to transform the value of the parameter before sending it to the server
  }>;
  response: ZodSchema; // you can use zod `transform` to transform the value of the response before returning it
  status?: number; // default to 200, you can use this to override the sucess status code of the response (only usefull for openapi and express)
  responseDescription?: string; // optional response description of the endpoint
  errors?: Array<{
    status: number | 'default';
    description?: string;
    schema: ZodSchema; // transformations are not supported on error schemas
  }>;
}>;

Full documentation

Check out the full documentation or following shortcuts.

Ecosystem

Roadmap for v11

for Zod/Io-Ts` :

  • By using the TypeProvider pattern we can now make zodios validation agnostic.

  • Implement at least ZodTypeProvider and IoTsTypeProvider since they both support input and output type inferrence

  • openapi generation will only be compatible with zod though

  • Not a breaking change so no codemod needed

  • MonoRepo:

    • Zodios will become a really large project so maybe migrate to turbo repo + pnpm

    • not a breaking change

  • Transform:

    • By default, activate transforms on backend and disable on frontend (today it's the opposite), would make server transform code simpler since with this option we could make any transforms activated not just zod defaults.

    • Rationale being that transformation can be viewed as business code that should be kept on backend

    • breaking change => codemod to keep current defaults by setting them explicitly

  • Axios:

    • Move Axios client to it's own package @zodios/axios and keep @zodios/core with only common types and helpers

    • Move plugins to @zodios/axios-plugins

    • breaking change => easy to do a codemod for this

  • Fetch:

    • Create a new Fetch client with almost the same features as axios, but without axios dependency @zodios/fetch

    • Today we have fetch support with a plugin for axios instance (zodios maintains it's own axios network adapter for fetch). But since axios interceptors are not used by zodios plugins, we can make fetch implementation lighter than axios instance.

    • Create plugins package @zodios/fetch-plugins

    • Not sure it's doable without a lot of effort to keep it in sync/compatible with axios client

    • new feature, so no codemod needed

  • React/Solid:

    • make ZodiosHooks independant of Zodios client instance (axios, fetch)

    • not a breaking change, so no codemod needed

  • Client Request Config

    • uniform Query/Mutation with body sent on the config and not as a standalone object. This would allow to not do client.deleteUser(undefined, { params: { id: 1 } }) but simply client.deleteUser({ params: { id: 1 } })

    • breaking change, so a codemod would be needed, but might be difficult to implement

  • Mock/Tests:

    • if we implement an abstraction layer for client instance, relying on moxios to mock APIs response will likely not work for fetch implementation.

    • create a @zodios/testing package that work for both axios/fetch clients

    • new feature, so no breaking change (no codemod needed)

You have other ideas ? Let me know !

Dependencies

Zodios even when working in pure Javascript is better suited to be working with Typescript Language Server to handle autocompletion. So you should at least use the one provided by your IDE (vscode integrates a typescript language server) However, we will only support fixing bugs related to typings for versions of Typescript Language v4.5 Earlier versions should work, but do not have TS tail recusion optimisation that impact the size of the API you can declare.

Also note that Zodios do not embed any dependency. It's your Job to install the peer dependencies you need.

Internally Zodios uses these libraries on all platforms :

  • zod
  • axios
项目侧边栏1项目侧边栏2
推荐项目
Project Cover

豆包MarsCode

豆包 MarsCode 是一款革命性的编程助手,通过AI技术提供代码补全、单测生成、代码解释和智能问答等功能,支持100+编程语言,与主流编辑器无缝集成,显著提升开发效率和代码质量。

Project Cover

AI写歌

Suno AI是一个革命性的AI音乐创作平台,能在短短30秒内帮助用户创作出一首完整的歌曲。无论是寻找创作灵感还是需要快速制作音乐,Suno AI都是音乐爱好者和专业人士的理想选择。

Project Cover

白日梦AI

白日梦AI提供专注于AI视频生成的多样化功能,包括文生视频、动态画面和形象生成等,帮助用户快速上手,创造专业级内容。

Project Cover

有言AI

有言平台提供一站式AIGC视频创作解决方案,通过智能技术简化视频制作流程。无论是企业宣传还是个人分享,有言都能帮助用户快速、轻松地制作出专业级别的视频内容。

Project Cover

Kimi

Kimi AI助手提供多语言对话支持,能够阅读和理解用户上传的文件内容,解析网页信息,并结合搜索结果为用户提供详尽的答案。无论是日常咨询还是专业问题,Kimi都能以友好、专业的方式提供帮助。

Project Cover

讯飞绘镜

讯飞绘镜是一个支持从创意到完整视频创作的智能平台,用户可以快速生成视频素材并创作独特的音乐视频和故事。平台提供多样化的主题和精选作品,帮助用户探索创意灵感。

Project Cover

讯飞文书

讯飞文书依托讯飞星火大模型,为文书写作者提供从素材筹备到稿件撰写及审稿的全程支持。通过录音智记和以稿写稿等功能,满足事务性工作的高频需求,帮助撰稿人节省精力,提高效率,优化工作与生活。

Project Cover

阿里绘蛙

绘蛙是阿里巴巴集团推出的革命性AI电商营销平台。利用尖端人工智能技术,为商家提供一键生成商品图和营销文案的服务,显著提升内容创作效率和营销效果。适用于淘宝、天猫等电商平台,让商品第一时间被种草。

Project Cover

AIWritePaper论文写作

AIWritePaper论文写作是一站式AI论文写作辅助工具,简化了选题、文献检索至论文撰写的整个过程。通过简单设定,平台可快速生成高质量论文大纲和全文,配合图表、参考文献等一应俱全,同时提供开题报告和答辩PPT等增值服务,保障数据安全,有效提升写作效率和论文质量。

投诉举报邮箱: service@vectorlightyear.com
@2024 懂AI·鲁ICP备2024100362号-6·鲁公网安备37021002001498号