Skip to content

gajus/graphql-lazyloader

master
Switch branches/tags

Name already in use

A tag already exists with the provided branch name. Many Git commands accept both tag and branch names, so creating this branch may cause unexpected behavior. Are you sure you want to create this branch?
Code

Latest commit

 

Git stats

Files

Permalink
Failed to load latest commit information.
Type
Name
Latest commit message
Commit time
src
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

graphql-lazyloader ๐Ÿ›‹

Travis build status Coveralls NPM version Canonical Code Style Twitter Follow

GraphQL directive that adds Object-level data resolvers.

Motivation

Several years ago I read GraphQL Resolvers: Best Practices (2018), an article written by PayPal team, that changed my view about where / when data resolution should happen.

Let's start with an example GraphQL schema:

type Query {
  person(id: ID) Person!
}

type Person {
  id: ID!
  givenName: String!
  familyName: String!
}

A typical GraphQL server uses "top-heavy" (parent-to-child) resolvers, i.e. in the above example, Query.person is responsible for fetching data for Person object. It may look something like this:

{
  Query: {
    person: (root, args) => {
      return getPerson(args.id);
    },
  },
};

PayPal team argues that this pattern is prone to data over-fetching. Instead, they propose to move data fetching logic to every field of Person, e.g.

{
  Query: {
    person: (root, args) => {
      return {
        id: args.id,
      };
    },
  },
  Person: {
    givenName: async ({id}) => {
      const {
        givenName,
      } = await getPerson(id);

      return givenName;
    },
    familyName: async ({id}) => {
      const {
        familyName,
      } = await getPerson(id);

      return givenName;
    },
  },
};

It is important to note that the above example assume that getPerson is implemented using a DataLoader pattern, i.e. data is fetched only once.

According to the original authors, this pattern is better because:

  • This code is easy to reason about. You know exactly where [givenName] is fetched. This makes for easy debugging.
  • This code is more testable. You don't have to test the [person] resolver when you really just wanted to test the [givenName] resolver.

To some, the [getPerson] duplication might look like a code smell. But, having code that is simple, easy to reason about, and is more testable is worth a little bit of duplication.

For this and other reasons, I became a fan โค๏ธ of this pattern and have since implemented it in multiple projects. However, the particular implementation proposed by PayPal is pretty verbose. graphql-lazyloader abstracts the above logic into a single GraphQL middleware (see Usage Example).

Usage

graphql-lazyloader is added using graphql-middleware

Usage Example

import {
  ApolloServer,
  gql,
} from 'apollo-server';
import {
  makeExecutableSchema,
} from '@graphql-tools/schema';
import {
  applyMiddleware,
} from 'graphql-middleware';
import {
  createLazyLoadMiddleware,
} from 'graphql-lazyloader';

const lazyLoadMiddleware = createLazyLoadMiddleware({
  Person: ({id}) => {
    return getPerson(id);
  },
});

const typeDefs = gql`
  type Query {
    person(id: ID!): Person!
  }

  type Person {
    id: ID!
    givenName: String!
    familyName: String!
  }
`;

const resolvers = {
  Query: {
    person: () => {
      return {
        id: '1',
      };
    },
  },
};

const schema = makeExecutableSchema({
  resolvers,
  typeDefs,
});

const schemaWithMiddleware = applyMiddleware(
  schema,
  lazyLoadMiddleware,
);

const server = new ApolloServer({
  schema: schemaWithMiddleware,
});

About

GraphQL directive that adds Object-level data resolvers.

Topics

Resources

License

Stars

Watchers

Forks