Skip to content

geek/graphi

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

graphi

hapi GraphQL server plugin

Build Status

Options

  • graphqlPath - HTTP path to serve graphql requests. Default is /graphql
  • graphiqlPath - HTTP path to serve the GraphiQL UI. Set to '' or false to disable. Default is /graphiql
  • schema - graphql schema either as a string or as a GraphQLSchema instance
  • resolvers - query and mutation functions mapped to their respective keys. Resolvers should return a promise when performing asynchronous operations.
  • authStrategy - (optional) Authentication strategy to apply to /graphql route. Default is false.
  • graphiAuthStrategy - (optional) Authentication strategy to apply to /graphiql route. Default is false.
  • formatError - (optional) Function that receives a GraphQLError as its only argument and returns a custom error object, which is returned to the client.
  • subscriptionOptions - (optional) Any options to pass to the nes subscription function.

API

The following decorations are made to the hapi server to make it easier to use a single graphi plugin with multiple other plugins depending on it.

  • server.registerSchema({ schema, resolvers }) - overwrites an existing registered schema with a new one.
  • server.makeExecutableSchema({ schema, resolvers, preResolve }) - combine resolvers with the schema definition into a GraphQLSchema.
  • server.plugins.graphi.publish(message, object) - Publish a message to any subscribers where message is the name of the message and object is the contents of the message.

The follow properties are exported directly when you require('graphi')

  • graphql - exported Graphql module that graphi uses
  • makeExecutableSchema({ schema, resolvers, preResolve }) - combine resolvers with the schema definition into a GraphQLSchema.

Events

The following server events are registered and available on server.events.on or server.events.once.

  • preFieldResolver - emitted before executing a resolver function. The event listener function is executed with an object argument that contains the following properties:

    • source
    • args
    • contextValue
    • info
  • postFieldResolver - emitted after executing a resolver function. The event listener function is executed with an object argument that contains the following properties:

    • source
    • args
    • contextValue
    • info
    • result

Usage

const schema = `
  type Person {
    firstname: String!
    lastname: String!
  }

  type Query {
    person(firstname: String!): Person!
  }
`;

const getPerson = function (args, request) {
  return { firstname: 'billy', lastname: 'jean' };
};

const resolvers = {
  person: getPerson
};

const server = Hapi.server();
await server.register({ plugin: Graphi, options: { schema, resolvers } });

With GraphQLSchema Instance

const schema = new GraphQLSchema({
  query: new GraphQLObjectType({
    name: 'RootQueryType',
    fields: {
      person: {
        type: GraphQLString,
        args: {
          firstname: { type: GraphQLString }
        },
        resolve: (root, { firstname }, request) => {
          return firstname;
        }
      }
    }
  })
});

const server = Hapi.server();
await server.register({ plugin: Graphi, options: { schema } });

With hapi routes

You can also define resolvers as hapi routes. As a result, each resolver is able to benefit from route caching, custom auth strategies, and all of the other powerful hapi routing features. Each route should either use the custom method 'graphql' or it should add a tag named 'graphql' and the path should be the key name for the resolver prefixed with /. You can also mix and match existing resolvers with routes.

const schema = `
  type Person {
    firstname: String!
    lastname: String!
  }

  type Query {
    person(firstname: String!): Person!
  }
`;


const server = Hapi.server();
server.route({
  method: 'graphql',
  path: '/person',
  handler: (request, h) => {
    // request.payload contains any arguments sent to the query
    return { firstname: 'billy', lastname: 'jean' };
  }
});

await server.register({ plugin: Graphi, options: { schema } });

This enables existing RESTful APIs to be easily converted over to GraphQL resolvers:

server.route({
  method: 'POST',
  path: '/person',
  config: {
    tags: ['graphql'],
    handler: (request, h) => {
      // request.payload contains any arguments sent to the query
      return { firstname: 'billy', lastname: 'jean' };
    }
  }
});

With Subscriptions

Graphi leverages nes to manage GraphQL subscriptions. Therefore, if you do intend to use subscriptions you will need to register nes with the hapi server. On the server a schema that contains subscriptions will automatically have those subscriptions registered with nes and graphi will expose helper functions to make publishing to subscribers easier. There is a server.plugins.graphi.publish(message, object) helper to make this easier to publish to any potential subscribers. Below is a complete example of registering a schema and then publishing to it.

const schema = `
  type Person {
    firstname: String!
    lastname: String!
    email: String!
  }

  type Subscription {
    personCreated(firstname: String!): Person!
  }
`;

const server = Hapi.server();
await server.register(Nes);
await server.register({ plugin: Graphi, options: { schema } });
await server.start();

server.plugins.graphi.publish('personCreated', { firstname: 'Peter', lastname: 'Pluck', email: 'test@test.com' });

Any clients that are subscribed to the personCreated event for the person with firstname = 'Peter' will receive the message that was published.

At the moment clients are required to use a nes compatible library and to subscribe to events using the client.subscribe function. The path that clients should use depends on the message, but in the previous example this would be '/personCreated/peter'.