Project Icon

SOFA

GraphQL到REST API的自动转换工具

SOFA是一个开源项目,可将GraphQL模式转换为RESTful API。它能分析GraphQL查询、变更和订阅,自动生成REST端点。该工具支持嵌套数据处理、自定义HTTP方法和路径、深度限制控制等功能。SOFA还能将GraphQL订阅转换为WebHook,并生成OpenAPI定义。这个工具为开发者提供了一种创建和管理API的方法,简化了GraphQL和REST API的集成过程。

GraphQLConf 2024 Banner: September 10-12, San Francisco. Hosted by the GraphQL Foundation

npm version Discord Chat code style: prettier renovate-app badge

The best way to create REST APIs (is GraphQL).

Installation

yarn add sofa-api
# or
npm install sofa-api

Getting Started

Here's complete example with no dependency on frameworks, but also integratable with any of them:

import http from 'http';
import getStream from 'get-stream';
import { useSofa } from 'sofa-api';

const server = http.createServer(
  useSofa({
    basePath: '/api',
    schema,
  })
);

Another example with builtin express-like frameworks support

import { useSofa } from 'sofa-api';
import express from 'express';

const app = express();

app.use(
  '/api',
  useSofa({
    basePath: '/api',
    schema,
  })
);

// GET /api/users
// GET /api/messages

How it works

Sofa takes your GraphQL Schema, looks for available queries, mutations and subscriptions and turns all of that into REST API.

Given the following schema:

type Chat {
  id: ID
  text: String
}

type Query {
  chat(id: ID): Chat
  chats: [Chat]
  me: Chat
  recentChats: [Chat]
}

Routes that are being generated:

GET /chat/:id
GET /chats
GET /me
GET /recent-chats

Nested data and idea behind Models

Sofa treats some types differently than others, those are called Models.

The idea behind Models is to not expose full objects in every response, especially if it's a nested, not first-level data.

For example, when fetching a list of chats you don't want to include all messages in the response, you want them to be just IDs (or links). Those messages would have to have their own endpoint. We call this type of data, a Model. In REST you probably call them Resources.

In order to treat particular types as Models you need to provide two queries, one that exposes a list (with no non-optional arguments) and the other to fetch a single entity (id field as an argument). The model itself has to have an id field. Those are the only requirements.

# Message is treated as a Model
type Query {
  messages: [Message]
  message(id: ID): Message
}

type Message {
  id: ID
  # other fields ...
}

Provide a Context

In order for Sofa to resolve operations based on a Context, you need te be able to provide some. Here's how you do it:

api.use(
  '/api',
  useSofa({
    basePath: '/api',
    schema,
    async context({ req }) {
      return {
        req,
        ...yourContext,
      };
    },
  })
);

You can pass a plain object or a function.

Use full responses instead of IDs

There are some cases where sending a full object makes more sense than passing only the ID. Sofa allows you to easily define where to ignore the default behavior:

api.use(
  '/api',
  useSofa({
    basePath: '/api',
    schema,
    ignore: ['Message.author'],
  })
);

Whenever Sofa tries to resolve an author of a message, instead of exposing an ID it will pass whole data.

Pattern is easy: Type.field or Type

Customize endpoint's HTTP Method, path and response status code

Sofa allows you to cutomize the http method, path and response status. For example, in case you need POST instead of GET method in one of your query, you do the following:

api.use(
  '/api',
  useSofa({
    schema,
    routes: {
      'Query.feed': { method: 'POST' },
    },
  })
);

When Sofa tries to define a route for feed of Query, instead of exposing it under GET (default for Query type) it will use POST method.

Pattern is easy: Type.field where Type is your query or mutation type.

You can also specify path with dynamic params support (for example /feed/:offset/:limit) and responseStatus.

Custom depth limit

Sofa prevents circular references by default, but only one level deep. In order to change it, set the depthLimit option to any number:

api.use(
  '/api',
  useSofa({
    basePath: '/api',
    schema,
    depthLimit: 2,
  })
);

Custom execute phase

By default, Sofa uses graphql function from graphql-js to turn an operation into data but it's very straightforward to pass your own logic. Thanks to that you can even use a remote GraphQL Server (with Fetch or through Apollo Links).

api.use(
  '/api',
  useSofa({
    basePath: '/api',
    schema,
    async execute(args) {
      return yourOwnLogicHere(args);
    },
  })
);

Subscriptions as webhooks

Sofa enables you to run GraphQL Subscriptions through WebHooks. It has a special API to start, update and stop a subscription.

  • POST /webhook - starts a subscription
  • DELETE /webhook/:id - stops it
  • POST /webhook/:id- updates it

Starting a subscription

To start a new subscription you need to include following data in request's body:

  • subscription - subscription's name, matches the name in GraphQL Schema
  • variables - variables passed to run a subscription (optional)
  • url - an url of your webhook receiving endpoint

After sending it to POST /webhook you're going to get in return a unique ID that is your started subscription's identifier.

{
  "id": "SUBSCRIPTION-UNIQUE-ID"
}

Stoping a subscription

In order to stop a subscription, you need to pass its id and hit DELETE /webhook/:id.

Updating a subscription

Updating a subscription looks very similar to how you start one. Your request's body should contain:

  • variables - variables passed to run a subscription (optional)

After sending it to POST /webhook/:id you're going to get in return a new ID:

{
  "id": "SUBSCRIPTION-UNIQUE-ID"
}

Example

Given the following schema:

type Subscription {
  onBook: Book
}

Let's start a subscription by sending that to POST /webhook:

{
  "subscription": "onBook",
  "variables": {},
  "url": "https://app.com/new-book"
}

In return we get an id that we later on use to stop or update subscription:

DELETE /webhook/:id

OpenAPI and Swagger

Thanks to GraphQL's Type System Sofa is able to generate OpenAPI (Swagger) definitions out of it. Possibilities are endless here. You get all the information you need in order to write your own definitions or create a plugin that follows any specification.

import { useSofa, OpenAPI } from 'sofa-api';
import { writeFileSync } from 'fs';

const openApi = OpenAPI({
  schema,
  info: {
    title: 'Example API',
    version: '3.0.0',
  },
});

app.use(
  '/api',
  useSofa({
    basePath: '/api',
    schema,
    onRoute(info) {
      openApi.addRoute(info, {
        basePath: '/api',
      });
    },
  })
);

// writes every recorder route
writeFileSync('./swagger.json', JSON.stringify(openApi.get(), null, 2));

OpenAPI (Swagger) with Bearer Authentication

import { useSofa, OpenAPI } from 'sofa-api';
import { writeFileSync } from 'fs';

const openApi = OpenAPI({
  schema,
  info: {
    title: 'Example API',
    version: '3.0.0',
  },
  components: {
    securitySchemes: {
      bearerAuth: {
        type: 'http',
        scheme: 'bearer',
        bearerFormat: 'JWT',
      },
    },
  },
  security: [
    {
      bearerAuth: [],
    },
  ],
});

app.use(
  '/api',
  useSofa({
    basePath: '/api',
    schema,
    onRoute(info) {
      openApi.addRoute(info, {
        basePath: '/api',
      });
    },
  })
);

// writes every recorder route
writeFileSync('./swagger.json', JSON.stringify(openApi.get(), null, 2));

OpenAPI (Swagger) with custom tags, summary and description

const openApi = OpenAPI({
  schema,
  info: {
    title: 'Example API',
    version: '3.0.0',
  },
  tags: [
    {
      name: 'Book',
      description: 'Book related operations',
    },
    {
      name: 'Author',
      description: 'Author related operations',
    },
  ],
});
@Resolver(Book)
export class BookResolver {
  @Query(() => Book, { description: 'Get book by id' }) // custom summary tag
  getBookById(@Arg('id', () => Int) id: number) {
    return 'book';
  }
}
const routes: SofaConfig['routes'] = {
  'Query.getBookById': {
    method: 'POST',
    path: '/book/:id',
    tags: ['Book'],
    description: 'This is a custom detailed description for getBookById',
  },
};

const createSofaMiddleware = (
  schema: GraphQLSchema,
  openApi: ReturnType<typeof OpenAPI>,
  basePath = ''
): ReturnType<typeof useSofa> => {
  return useSofa({
    routes,
    basePath,
    schema,
    onRoute(info) {
      openApi.addRoute(info, { basePath });
    },
  });
};
// writes every recorder route
openApi.save('./swagger.yml');

License

MIT © Uri Goldshtein

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