GraphQL Context

You can use your configuration to populate the GraphQL context based on the wider context in which the GraphQL request is being executed.

Making HTTP data available to plan resolvers

Using the grafast.context callback we can extract data from the incoming HTTP request and make it accessible from within the Grafast schema via the GraphQL context.

Example:

graphile.config.js

export default {
  grafast: {
    async context(requestContext, args) {
      const req = requestContext.node?.req;
      // You can perform asynchronous actions here if you need to; for example
      // looking up the current user in the database.

      // Return here things that your resolvers need
      return {
        // Return the current user from Passport.js or similar
        user: req.user,

        // Add a helper to get a header
        getHeader(name) {
          return req?.get(name);
        },

        // Give access to the database-owner PostgreSQL pool, for example to
        // perform privileged actions
        rootPgPool,
      };
    },
  },
};

Use prefixes

When adding details to grafast.context, you must careful to not add properties that will clash with system context keys such as withPgClient, pgSettings, pgSubscriber and jwtClaims (you can see the existing context keys by inspecting the contextValue property of second argument to the grafast.context callback: args.contextValue).

For the absolute best future compatibility, we recommend that you prefix your context keys with your initials, company name, or similar.

Use helpers

It’s not a good idea to give direct access to the req or res objects via grafast.context as it binds the GraphQL context too tightly to the HTTP request lifecycle — this will cause you issues if you try and use the GraphQL schema in other contexts (e.g. directly from the application, in integration tests, or over alternative transports such as websockets for realtime). Instead, add helpers to get/set the data you need that can be implemented in each future situation.