​
Subgraph object

The subgraph object is the core of the composition library. It has the following form:

import { DocumentNode } from 'graphql';
import { federateSubgraphs, Subgraph } from '@wundergraph/composition';

export type Subgraph = {
  definitions: DocumentNode
  name: string;
  url: string;
};

A DocumentNode can be created using the graphql libary parse function. A simple example subgraph is provided below :

import { DocumentNode } from 'graphql';
import { federateSubgraphs, Subgraph } from '@wundergraph/composition';

const subgraphA = {
  name: 'subgraph-a',
  url: 'http://localhost:4001',
  definitions: parse(`
    type User {
      name: String!
    }
  `),
};

​
Federating subgraphs

Producing a federating graph with WunderMerge.

​
federateSubgraphs

The federateSubgraphs function is responsible for normalizing, validating, and finally federating an array of Subgraph objects:

export function federateSubgraphs(subgraphs: Subgraph[]): FederationResult;

​
Federation Result

The federateSubgraphs function returns a FederationResult, which has the following form:

import { DocumentNode, GraphQLSchema } from 'graphql';

export type FederationResult = {
  errors?: Error[];
  federatedGraphAST?: DocumentNode;
  federatedGraphSchema?: GraphQLSchema;
};

​
Properties

errors: An optional array of all errors encountered during normalization and validation of the subgraph, or composition and validation of the federated graph. If there are any errors during normalization, federation will not be attempted. If federation was successful, the array will be undefined.

federatedGraphAST: An optional graphql DocumentNode representation of the federated graph. If federation was unsuccessful, this property will be undefined.

federatedGraphSchema: An optional graphql GraphQLSchema representation of the federated graph. If federation was unsuccessful, this property will be undefined.

​
Printing the federated schema

You can produce a string of the federated graph through graphql print or graphql printSchema.

import { print, printSchema } from 'graphql';

const result = federateSubgraphs([subgraphA, subgraphB]);
if (result.errors) {
  for (const err of result.errors) {
    console.log(err.message);
  }
} else {
  print(result.federatedGraphAST!);
  printSchema(result.federatedGraphSchema!);
}

​
Example

An example of federation with two graphs.

# subgraph 1
type Query {
  employees: [Employee!]!
  employee(id: Int!): Employee
}

enum Department {
  ENGINEERING
  MARKETING
  OPERATIONS
}

interface RoleType {
  department: Department!
}

enum EngineerType {
  FRONTEND
  BACKEND
  FULLSTACK
}

type Engineer implements RoleType {
  department: Department!
  type: EngineerType!
}

type Marketer implements RoleType {
  department: Department!
}

type Operator implements RoleType {
  department: Department!
}

type Employee @key(fields: "id") {
  id: Int!
  forename: String!
  surname: String!
  role: RoleType!
}

# subgraph 2
enum ProductNames {
  CLOUD
  COSMO
  ENGINE
  SDK
}

type Employee @key(fields: "id") {
  id: Int!
  products: [ProductNames!]
}

import { parse, print } from 'graphql';

const subgraphOne = {
  name: 'subgraph-1',
  url: 'http://localhost:4001',
  definitions: parse(subgraphOneSDLString),
}: Subgraph;

const subgraphTwo = {
  name: 'subgraph-2',
  url: 'http://localhost:4002',
  definitions: parse(subgraphTwoSDLString),
}: Subgraph;

const result = federateSubgraphs([subgraphOne, subgraphTwo]);
if (result.errors) {
  for (const err of result.errors) {
    console.log(err.message);
  }
} else {
  print(result.federatedGraphAST!);
}

# federated graph
directive @deprecated(reason: String = "No longer supported") on ARGUMENT_DEFINITION | ENUM_VALUE | FIELD_DEFINITION | INPUT_FIELD_DEFINITION
directive @external on FIELD_DEFINITION | OBJECT
directive @key(fields: String!) repeatable on OBJECT
directive @provides(fields: String!) on FIELD_DEFINITION
directive @requires(fields: String!) on FIELD_DEFINITION
directive @tag(name: String!) repeatable on ARGUMENT_DEFINITION | ENUM | ENUM_VALUE | FIELD_DEFINITION | INPUT_FIELD_DEFINITION | INPUT_OBJECT | INTERFACE | OBJECT | SCALAR | UNION

type Query {
  employees: [Employee!]!
  employee(id: Int!): Employee
}

enum Department {
  ENGINEERING
  MARKETING
  OPERATIONS
}

interface RoleType {
  department: Department!
}

enum EngineerType {
  FRONTEND
  BACKEND
  FULLSTACK
}

type Engineer implements RoleType {
  department: Department!
  type: EngineerType!
}

type Marketer implements RoleType {
  department: Department!
}

type Operator implements RoleType {
  department: Department!
}

type Employee {
  id: Int!
  forename: String!
  surname: String!
  role: RoleType!
  products: [ProductNames!]
}

enum ProductNames {
  CLOUD
  COSMO
  ENGINE
  SDK
}

Was this page helpful?