This project is still maintained but has been superseded by graphql-schemax. graphql-schemax takes the approach of compiling standard JSON object into a GraphQL Schema string.
GraphQL S2S enriches the standard GraphQL Schema string used by both graphql.js and the Apollo Server. The enriched schema supports:
- Type Inheritance
- Generic Types
- Metadata Decoration
- Deconstructing - Transforming - Rebuilding Queries
npm install graphql-s2s --save
<script src="https://unpkg.com/graphql-s2s@0.11.2/lib/graphqls2s.min.js"></script>
Using the awesome unpkg.com, all versions are supported at https://unpkg.com/graphql-s2s@:VERSION/lib/graphqls2s.min.js. The API will be accessible through the graphqls2s object.
It is also possible to embed it after installing the graphql-s2s npm package:
<script src="./node_modules/graphql-s2s/lib/graphqls2s.min.js"></script>
const { transpileSchema } = require('graphql-s2s').graphqls2s
const { makeExecutableSchema } = require('graphql-tools')
const schema = `
type Node {
id: ID!
}
type Person inherits Node {
firstname: String
lastname: String
}
type Student inherits Person {
nickname: String
}
type Query {
students: [Student]
}
`
const resolver = {
Query: {
students(root, args, context) {
// replace this dummy code with your own logic to extract students.
return [{ id: 1, firstname: "Carry", lastname: "Connor", nickname: "Cannie" }]
}
}
};
const executableSchema = makeExecutableSchema({
typeDefs: [transpileSchema(schema)],
resolvers: resolver
})
const schema = `
type Node {
id: ID!
}
# Inheriting from the 'Node' type
type Person inherits Node {
firstname: String
lastname: String
}
# Inheriting from the 'Person' type
type Student inherits Person {
nickname: String
}
`
const schema = `
type Node {
id: ID!
}
type Address {
streetAddress: String
city: String
state: String
}
# Inheriting from the 'Node' & 'Adress' type
type Person inherits Node, Address {
id: ID!
streetAddress: String
city: String
state: String
firstname: String
lastname: String
}
`
More details in the code below.
const schema = `
# Defining a generic type
type Paged<T> {
data: [T]
cursor: ID
}
type Question {
name: String!
text: String!
}
# Using the generic type
type Student {
name: String
questions: Paged<Question>
}
# Using the generic type
type Teacher {
name: String
students: Paged<Student>
}
`
More details in the code below.
const schema = `
# Defining a custom 'node' metadata attribute
@node
type Node {
id: ID!
}
type Student inherits Node {
name: String
# Defining another custom 'edge' metadata, and supporting a generic type
@edge(some other metadata using whatever syntax I want)
questions: [String]
}
`
The enriched schema provides a richer and more compact notation. The transpiler converts the enriched schema into the standard expected by graphql.js (using the buildSchema method) as well as the Apollo Server. For more details on how to extract those extra information from the string schema, use the method getSchemaAST (example in section Metadata Decoration).
Metadata can be added to decorate the schema types and properties. Add whatever you want as long as it starts with @ and start hacking your schema. The original intent of that feature was to decorate the schema with metadata @node and @edge so we could add metadata about the nature of the relations between types.
Metadata can also be used to customize generic types names as shown in section How to use a custom name on generic types?.
This feature allows your GraphQl server to deconstruct any GraphQl query as an AST that can then be filtered and modified based on your requirements. That AST can then be rebuilt as a valid GraphQL query. A great example of that feature in action is the graphql-authorize middleware for graphql-serverless which filters query's properties based on the user's rights.
For a concrete example, refer to the code below.
Use the special keyword @alias
as follow:
const schema = `
type Post {
code: String
}
type Brand {
id: ID!
name: String
posts: Page<Post>
}
@alias((T) => T + 's')
type Page<T> {
data: [T]
}
`
After transpilation, the resulting schema is:
const output = transpileSchema(schema)
// output:
// =======
// type Post {
// code: String
// }
//
// type Brand {
// id: ID!
// name: String
// posts: Posts
// }
//
// type Posts {
// data: [Post]
// }
WARNING: the following examples will be based on 'graphql-tools' from the Apollo team, but the string schema could also be used with the 'buildSchema' method from graphql.js
NOTE: The examples below only use 'type', but it would also work on 'input' and 'interface'
Before graphql-s2s
const schema = `
type Teacher {
id: ID!
creationDate: String
firstname: String!
middlename: String
lastname: String!
age: Int!
gender: String
title: String!
}
type Student {
id: ID!
creationDate: String
firstname: String!
middlename: String
lastname: String!
age: Int!
gender: String
nickname: String!
}`
After graphql-s2s
const schema = `
type Node {
id: ID!
creationDate: String
}
type Person inherits Node {
firstname: String!
middlename: String
lastname: String!
age: Int!
gender: String
}
type Teacher inherits Person {
title: String!
}
type Student inherits Person {
nickname: String!
}`
Full code example
const { transpileSchema } = require('graphql-s2s').graphqls2s
const { makeExecutableSchema } = require('graphql-tools')
const { students, teachers } = require('./dummydata.json')
const schema = `
type Node {
id: ID!
creationDate: String
}
type Person inherits Node {
firstname: String!
middlename: String
lastname: String!
age: Int!
gender: String
}
type Teacher inherits Person {
title: String!
}
type Student inherits Person {
nickname: String!
questions: [Question]
}
type Question inherits Node {
name: String!
text: String!
}
type Query {
# ### GET all users
#
students: [Student]
# ### GET all teachers
#
teachers: [Teacher]
}
`
const resolver = {
Query: {
students(root, args, context) {
return Promise.resolve(students)
},
teachers(root, args, context) {
return Promise.resolve(teachers)
}
}
}
const executableSchema = makeExecutableSchema({
typeDefs: [transpileSchema(schema)],
resolvers: resolver
})
NOTE: The examples below only use 'type', but it would also work on 'input'
Before graphql-s2s
const schema = `
type Teacher {
id: ID!
creationDate: String
firstname: String!
middlename: String
lastname: String!
age: Int!
gender: String
title: String!
}
type Student {
id: ID!
creationDate: String
firstname: String!
middlename: String
lastname: String!
age: Int!
gender: String
nickname: String!
questions: Questions
}
type Question {
id: ID!
creationDate: String
name: String!
text: String!
}
type Teachers {
data: [Teacher]
cursor: ID
}
type Students {
data: [Student]
cursor: ID
}
type Questions {
data: [Question]
cursor: ID
}
type Query {
# ### GET all users
#
students: Students
# ### GET all teachers
#
teachers: Teachers
}
`
After graphql-s2s
const schema = `
type Paged<T> {
data: [T]
cursor: ID
}
type Node {
id: ID!
creationDate: String
}
type Person inherits Node {
firstname: String!
middlename: String
lastname: String!
age: Int!
gender: String
}
type Teacher inherits Person {
title: String!
}
type Student inherits Person {
nickname: String!
questions: Paged<Question>
}
type Question inherits Node {
name: String!
text: String!
}
input Filter<FilterFields> {
field: FilterFields!,
value: String!
}
enum TeachersFilterFields {
firstName
lastName
}
type Query {
# ### GET all users
#
students: Paged<Student>
# ### GET all teachers
# You can use generic types on parameters, too.
#
teachers(filter: Filter<TeachersFilterFields>): Paged<Teacher>
}
`
This is very similar to C# or Java generic classes. What the transpiler will do is to simply recreate 3 types (one for Paged<Question>, Paged<Student> and Paged<Teacher>), and one input (Filter<TeachersFilterFields>). If we take the Paged<Question> example, the transpiled type will be:
type PagedQuestion {
data: [Question]
cursor: ID
}
Full code example
const { transpileSchema } = require('graphql-s2s').graphqls2s
const { makeExecutableSchema } = require('graphql-tools')
const { students, teachers } = require('./dummydata.json')
const schema = `
type Paged<T> {
data: [T]
cursor: ID
}
type Node {
id: ID!
creationDate: String
}
type Person inherits Node {
firstname: String!
middlename: String
lastname: String!
age: Int!
gender: String
}
type Teacher inherits Person {
title: String!
}
type Student inherits Person {
nickname: String!
questions: Paged<Question>
}
type Question inherits Node {
name: String!
text: String!
}
type Query {
# ### GET all users
#
students: Paged<Student>
# ### GET all teachers
#
teachers: Paged<Teacher>
}
`
const resolver = {
Query: {
students(root, args, context) {
return Promise.resolve({ data: students.map(s => ({ __proto__:s, questions: { data: s.questions, cursor: null }})), cursor: null })
},
teachers(root, args, context) {
return Promise.resolve({ data: teachers, cursor: null });
}
}
}
const executableSchema = makeExecutableSchema({
typeDefs: [transpileSchema(schema)],
resolvers: resolver
})
Define your own custom metadata and decorate your GraphQL schema with new types of data. Let's imagine we want to explicitely add metadata about the type of relations between nodes, we could write something like this:
const { getSchemaAST } = require('graphql-s2s').graphqls2s
const schema = `
@node
type User {
@edge('<-[CREATEDBY]-')
posts: [Post]
}
`
const schemaObjects = getSchemaAST(schema);
// -> schemaObjects
// {
// "type": "TYPE",
// "name": "User",
// "metadata": {
// "name": "node",
// "body": "",
// "schemaType": "TYPE",
// "schemaName": "User", "parent": null
// },
// "genericType": null,
// "blockProps": [{
// "comments": "",
// "details": {
// "name": "posts",
// "metadata": {
// "name": "edge",
// "body": "(\'<-[CREATEDBY]-\')",
// "schemaType": "PROPERTY",
// "schemaName": "posts: [Post]",
// "parent": {
// "type": "TYPE",
// "name": "User",
// "metadata": {
// "type": "TYPE",
// "name": "node"
// }
// }
// },
// "params": null,
// "result": {
// "originName": "[Post]",
// "isGen": false,
// "name": "[Post]"
// }
// },
// "value": "posts: [Post]"
// }],
// "inherits": null,
// "implements": null
// }
This feature allows your GraphQl server to deconstruct any GraphQl query as an AST that can then be filtered and modified based on your requirements. That AST can then be rebuilt as a valid GraphQL query. A great example of that feature in action is the graphql-authorize middleware for graphql-serverless which filters query's properties based on the user's rights.
const { getQueryAST, buildQuery, getSchemaAST } = require('graphql-s2s').graphqls2s
const schema = `
type Property {
name: String
@auth
address: String
}
input InputWhere {
name: String
locations: [LocationInput]
}
input LocationInput {
type: String
value: String
}
type Query {
properties(where: InputWhere): [Property]
}`
const query = `
query {
properties(where: { name: "Love", locations: [{ type: "house", value: "Bellevue hill" }] }){
name
address
}
}`
const schemaAST = getSchemaAST(schema)
const queryAST = getQueryAST(query, null, schemaAST)
const rebuiltQuery = buildQuery(queryAST.filter(x => !x.metadata || x.metadata.name != 'auth'))
// query {
// properties(where:{name:"Love",locations:[{type:"house",value:"Bellevue hill"}]}){
// name
// }
// }
Notice that the original query was requesting the address
property. Because we decorated that property with the custom metadata @auth
(feature demonstrated previously Metadata Decoration), we were able to filter that property to then rebuilt the query without it.
getQueryAST(query, operationName, schemaAST, options): QueryAST
Returns an GraphQl query AST.
Arguments | type | Description |
---|---|---|
query | String | GraphQl Query. |
operationName | String | GraphQl query operation. Only useful if multiple operations are defined in a single query, otherwise use null . |
schemaAST | Object | Original GraphQl schema AST obtained thanks to the getSchemaAST function. |
options.defrag | Boolean | If set to true and if the query contained fragments, then all fragments are replaced by their explicit definition in the AST. |
QueryAST Object Structure
Properties | type | Description |
---|---|---|
name | String | Field's name. |
kind | String | Field's kind. |
type | String | Field's type. |
metadata | String | Field's metadata. |
args | Array | Array of argument objects. |
properties | Array | Array of QueryAST objects. |
QueryAST.filter((ast:QueryAST) => ...): QueryAST
Returns a new QueryAST object where only ASTs complying to the predicate ast => ...
are left.
QueryAST.propertyPaths((ast:QueryAST) => ...): [String]
Returns an array of strings. Each one represents the path to the query property that matches the predicate ast => ...
.
QueryAST.containsProp(property:String): Boolean
Returns a boolean indicating the presence of a property in the GraphQl query. Example:
const schema = `
type User {
id: ID!
name: String
details: UserDetails
}
type UserDetails {
gender: String
}
type Query {
users: [User]
}
`
const query = `
{
users {
id
details {
gender
}
}
}`
const schemaAST = getSchemaAST(schema)
const queryAST = getQueryAST(query, null, schemaAST)
queryAST.containsProp('users.id') // true
queryAST.containsProp('users.details.gender') // true
queryAST.containsProp('details.gender') // true
queryAST.containsProp('users.name') // false
QueryAST.some((ast:QueryAST) => ...): Boolean
Returns a boolean indicating whether the QueryAST contains at least one AST matching the predicate ast => ...
.
buildQuery(ast:QueryAST): String
Rebuilds a valid GraphQl query from a QueryAST object.
We only accept pull request that have been thoroughly tested. To do so, simply add your test under the test/browser/graphqls2s.js
file.
Once that's done, simply run your the following command to test your features:
npm run test:dev
This sets an environment variable that configure the project to load the main dependency from the src folder (source code in ES6) instead of the lib folder (transpiled ES5 code).
npm run dev
npm run build
npm test
This project is built using Javascript ES6. Each version is also transpiled to ES5 using Babel through Webpack 2, so this project can run in the browser. In order to write unit test only once instead of duplicating it for each version of Javascript, the all unit tests have been written using Javascript ES5 in mocha. That means that if you want to test the project after some changes, you will need to first transpile the project to ES5.
We are Neap, an Australian Technology consultancy powering the startup ecosystem in Sydney. We simply love building Tech and also meeting new people, so don't hesitate to connect with us at https://neap.co.
Our other open-sourced projects:
- graphql-s2s: Add GraphQL Schema support for type inheritance, generic typing, metadata decoration. Transpile the enriched GraphQL string schema into the standard string schema understood by graphql.js and the Apollo server client.
- schemaglue: Naturally breaks down your monolithic graphql schema into bits and pieces and then glue them back together.
- graphql-authorize: Authorization middleware for graphql-serverless. Add inline authorization straight into your GraphQl schema to restrict access to certain fields based on your user's rights.
- react-native-game-engine: A lightweight game engine for react native.
- react-native-game-engine-handbook: A React Native app showcasing some examples using react-native-game-engine.
- core-async: JS implementation of the Clojure core.async library aimed at implementing CSP (Concurrent Sequential Process) programming style. Designed to be used with the npm package 'co'.
- jwt-pwd: Tiny encryption helper to manage JWT tokens and encrypt and validate passwords using methods such as md5, sha1, sha256, sha512, ripemd160.
- google-cloud-bucket: Nodejs package to manage Google Cloud Buckets and perform CRUD operations against them.
- google-cloud-bigquery: Nodejs package to manage Google Cloud BigQuery datasets, and tables and perform CRUD operations against them.
- google-cloud-tasks: Nodejs package to push tasks to Google Cloud Tasks. Include pushing batches.
Copyright (c) 2017-2019, Neap Pty Ltd. All rights reserved.
Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met:
- Redistributions of source code must retain the above copyright notice, this list of conditions and the following disclaimer.
- Redistributions in binary form must reproduce the above copyright notice, this list of conditions and the following disclaimer in the documentation and/or other materials provided with the distribution.
- Neither the name of Neap Pty Ltd nor the names of its contributors may be used to endorse or promote products derived from this software without specific prior written permission.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL NEAP PTY LTD BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.