Code for the talk on VanJS meetup about GraphQL.
- npm
- git
- node
To set up the code you will need to pull down the source codes from github repo as follows:
cd ~
git clone https://github.com/argemiront/vanjs-graphql.git
cd vanjs-graphql
npm installYou can serve the API at your local machine on http://localhost:3000. To start serving the API,
npm startYou can change the port on the index.js from the root folder of the project in the following line:
server.listen({ port: 3000 }).then(({ url }) => console.log(`serving at: ${url}`));
Basic structure with schema, resolvers and server inside the index.js file.
The graphql folder is created. Schemas and its resolvers now are in separated folders.
The merge-graphql-schemas library is used to automatically merge schemas and resolvers inside the folder that follows the pattern:
- schemas: schema_name.typedefs.gql
- resolvers: resolver_name.resolvers.js
The data is moved to the database.js lib inside the utils folder. It provides functions to fetch the data.
Middleware are added to the project in 3 different ways:
- added in the GraphQL context for authentication purpose;
- to run with all resolvers;
- to run before a specific resolver;
Middleware are merged with the original schema and resolvers by the applyMiddleware functions from the graphql-middleware library.
The dataloader middleware is added here to batch and cache database requests.
The branch has two parts, on 4.1 it is only added a console.log on the database library to demonstrate the issue with multiple data load. On 4.2 the issue is solved by applying the dataloader.
In the last part two directives are created to perform additional computation on the server before the data is sent to the client.
Creating a new directive can be accomplished with the following procedures:
- add the directive to the schema
- extend the SchemaDirectiveVisitor class
- override the desired visitor method