README

dotql

graphql like but query by plain json. No more DSL or AST parsing

npm npm npm

A graphql like client and server without graphql DSL, faster parsing.

Built-in implementation for live-query (by ping or/and push)

List of features

strictly follow graphql but in json, no more DSL parsing
small in size, all-in-one instead of few heavy weighted packages
built-in implementation for live-query (by ping or/and push)

Basic Usage

dotql query

client.query({
  getUserById: {
    // with id:'user_01' arguments
    $args: { id: 'user_01' },
    // ask for name and photo fields
    name: 1,
    photo: 1,
  },
})

instead of graphql query

graphqlClient.query(gql`
  {
    getUserById(id: "user_01") {
      name
      photo
    }
  }
`)

mutate and watch server data via client

const client = new Client({
  callServer: specs => server.query(specs),
})

// watch
const unwatch = client.watch({ getUserById: { $args: 'user_01', id: 1, count: 1 } }, (data, error) => {
  watchData = data
})
await delay()
// watch will fill initial data
expect(watchData).toMatchObject({ getUserById: { $type: 'User', id: 'user_01', count: undefined } })

// mutate
await client.mutate({ setUserById: { $args: { id: 'user_01', count: 10 } } })
await delay()
expect(watchData).toMatchObject({ getUserById: { $type: 'User', id: 'user_01', count: 10 } })

use prepared queries for smaller request payload and hiding schema detail

// use prepared query 'getUserById_1'
expect(await client.query({ $query: 'getUserById_1', where: 'user_01' })).toMatchObject({
  getUserById: { $type: 'User', id: 'user_01' },
})

// use prepared mutation 'setUserById_1'
await client.mutate({ $mutation: 'setUserById_1', userId: 'user_01', count: 20 })
await delay()
expect(watchData).toMatchObject({ getUserById: { $type: 'User', id: 'user_01', count: 20 } })

create server instance

const server = new Server({
  schema: {
    Queries: {
      getUserById: {
        type: 'User',
        resolve: async (dot, args, context, info) => {
          return { ...userDb[args], id: args }
        },
      },
    },
    Mutations: {
      setUserById: {
        type: 'User',
        resolve: async (dot, args, context, info) => {
          return (userDb[args.id] = Object.assign(userDb[args.id] || {}, args))
        },
      },
    },
    User: {
      id: { type: 'String' },
      count: { type: 'Int' },
      blogs: {
        type: ['Object'],
        resolve: async (dot, args, context, info) => {
          // batch.load() will auto create new DataLoader(batchLoader)
          const blogId = await context.batch.load(dot.id)
          return [{ blogId }]
        },
        batchLoader: async keys => _.map(keys, k => `< ${k} >`),
      },
    },
  },
  prepared: {
    Queries: {
      // prepared query with name='getUserById_1'
      getUserById_1: {
        // { $ref: 'where' } will be replace by prepared query references
        getUserById: { $args: { $ref: 'where' }, id: 1, count: 1 },
      },
    },
    Mutations: {
      setUserById_1: {
        // $refs will pick multiple variables
        setUserById: { $args: { $refs: { id: 'userId', count: 'count' } } },
      },
    },
  },
})

trigger resolve to call context.batch.load() which use built-in support of dataloader

// resolve via batchLoader
result = await client.query({ getUserById: { $args: 'user_01', blogs: 1 } })
expect(result).toMatchObject({ getUserById: { $type: 'User', blogs: [{ blogId: '< user_01 >' }] } })

TODO

Docs

eTags and channels
useMutate(func, deps)

Size

dotql client and server

npm

graphql (not yet include apollo or relay)

npm

Contributing

Keep it simple. Keep it minimal. Don't put every single feature just because you can.

Usage no npm install needed!