apollo-link-webworker

Apollo link that lets you use graphql client-side only, with a webworker as a "server" supporting normal query and subscriptions

Important note

This repository is just a proof of concept and not intended for production use yet. But contributions are welcomed :)

Installing

Install the package and its peer dependencies : yarn add apollo-link-webworker graphql apollo-link subscriptions-transport-ws

Getting started

Start by creating a worker.js file. Then you can import the utility functions that help you to build the worker :

worker.js

import { createWorker, handleSubscriptions } from 'apollo-link-webworker';

Creating the basic worker

createWorker takes an option object as parameter accepting the schema and the context:

worker.js

import { createWorker, handleSubscriptions } from 'apollo-link-webworker';

import schema from './schema'; // your graphql schema
import context from './context'; // your graphql context

createWorker({
  schema,
  context
});

Configuring the webpack worker-loader

In order to require the worker file, you'll need to add the worker-loader to your webpack configuration :

yarn add worker-loader --dev

Then, with an ejected react-app for example, edit the config/webpack.config.[dev/prod].js files to add the specific loader :

[...]
module: {
    strictExportPresence: true,
    rules: [
      // TODO: Disable require.ensure as it's not a standard language feature.
      // We are waiting for https://github.com/facebookincubator/create-react-app/issues/2176.
      // { parser: { requireEnsure: false } },

      // First, run the linter.
      // It's important to do this before Babel processes the JS.
      {
        test: /\.(js|jsx|mjs)$/,
        enforce: 'pre',
        use: [
          {
            options: {
              formatter: eslintFormatter,
              eslintPath: require.resolve('eslint'),
              
            },
            loader: require.resolve('eslint-loader'),
          },
        ],
        include: paths.appSrc,
      },
      {
        // "oneOf" will traverse all following loaders until one will
        // match the requirements. When no loader matches it will fall
        // back to the "file" loader at the end of the loader list.
        oneOf: [
          {
            test: /worker\.js$/,  //worker.js is the filename I chose
            include: path.appSrc,
            loader: require.resolve('worker-loader'),
          },
          // "url" loader works like "file" loader except that it embeds assets
          // smaller than specified limit in bytes as data URLs to avoid requests.
          // A missing `test` is equivalent to a match.
          {
            test: [/\.bmp$/, /\.gif$/, /\.jpe?g$/, /\.png$/],
            loader: require.resolve('url-loader'),
            options: {
              limit: 10000,
              name: 'static/media/[name].[hash:8].[ext]',
            },
          },
[...]

If you only want to support classical communication (i.e : you don't mind about subscriptions) you can skip the next step.

Handling subscriptions

apollo-link-webworker lets you generate graphql subscriptions from external event source. It's very useful when you don't own the socket server (firebase realtime database for example) but still need to resolve the result via your graphql schema.

To add subscriptions, just compose the handleSubscriptions utility function inside the worker onmessage handler :

worker.js

import { createWorker, handleSubscriptions } from 'apollo-link-webworker';

import schema from './schema'; // your graphql schema
import context from './context'; // your graphql context
import pubsub from './pubsub'; // a PubSub instance from graphql-subscriptions package for example

createWorker({
  schema,
  context
});

self.onmessage = message => handleSubscriptions({
  self,
  message,
  schema,
  context,
  pubsub,
});

pubsub.js

import { PubSub } from 'graphql-subscriptions';

const pubsub = new PubSub();

export default pubsub;

Whenever you need to generate a subscription from an external event source, you just need to push to the correct pubsub channel. By convention, the channel name used is the graphql subscription operation name.

For example (with firebase) : schema.js

import pubsub from './pubsub';

const schemaString = `
  [...]

  type Subscription {
    messageAdded: Message!
  }

  [...]

`;

const resolvers = {
  [...]
  Subscription: {
    messageAdded: {
      subscribe: () => pubsub.asyncIterator('OnMessageAdded'),
    },
  }
  [...]
};

OnMessageAdded.graphql

subscription OnMessageAdded {
  messageAdded {
    id
    content
    user {
      id
      username
    }
  }
}

// Generates a subscriptions from external firebase event source
firebaseDb().ref('/messages').on('child_added', snapshot => pubsub.publish('OnMessageAdded', {
 messageAdded: snapshot.val()
}));

Generating the apollo client

Once you created your worker.js file you can instanciate a new WebWorkerLink from the factory function createWebWorkerLink :

client.js

import { ApolloClient } from 'apollo-client';
import InMemoryCache from 'apollo-cache-inmemory';
import { createWebWorkerLink } from 'apollo-link-webworker';

const GraphqlWorker = require('./worker.js');

const worker = new GraphqlWorker();

const link = createWebWorkerLink({ worker });

const dataIdFromObject = result => result.id;

const cache = new InMemoryCache({ dataIdFromObject });

const client = new ApolloClient({
  cache,
  link,
});

export default client;

Example chat application with Firebase & Authentication :

Firechat repository