Setting up a graphql api for #52Projects
- Clone
- Run
yarn && yarn start - Go to
http://localhost:4000
Uses the file system to generate a list of projects
Examples:
query {
projects {
name
id
type
}
}
query {
project() {
name
id
type
}
}
query {
project(input: {name: "Hangman React"}) {
name
id
type
}
}
query {
projects(input: {type: FRONTEND}) {
name
id
type
}
}
mutation {
createProject(input: {name: "new project", number: 5, type: BACKEND}) {
name
id
type
number
}
}
We're implementing the Project interface to tell graphql that both OngoingProject and CompleteProject are of type Project. Using interface allows us to use the implements keyword and tell graphql both structures refer to the same object with slight changes. Note that we still need to copy the common fields (name, id, etc).
interface Project {
name: String!
id: String!
number: Int!
repo: String
url: String
type: ProjectType
}
type OngoingProject implements Project {
name: String!
id: String!
number: Int!
repo: String
url: String
type: ProjectType
isTested: Boolean!
todo: [String]!
}
type CompleteProject implements Project {
name: String!
id: String!
number: Int!
repo: String
url: String
type: ProjectType
isComplete: Boolean!
}
Once these types and interface are in the schema, if we try to start the server - we'll get an error.
Type "Project" is missing a "__resolveType" resolver. Pass false into "resolverValidationOptions.requireResolversForResolveType" to disable this warning.
We need to allow graphql to distinguish between the two implementations of Projects. In our example, we have a todo array on every ongoing project. The resolver might look like this:
Project: {
__resolveType(project) {
return (project.todo && project.todo.length > 0 ? 'OngoingProject' : 'CompleteProject')
}
}
This next query gets all commong fields from all projects, and also gets the todo fields from the OngoingProject items.
query {
projects {
name
id
type
number
... on OngoingProject {
todo
}
}
}