graphql-schema-builder

Builds GraphQL types based on Mongoose-like schema and a resolver list thus separating type schema from resolvers. Simplifies defining mutation arguments based on schema types.

Type's fields can be used to build Mongoose schema by calling it with mongoose.Schema.Types or mechanically converted into any other schemas.

NB: All types are assumed to have an implicit id field.

Usage

yarn add graphql-schema-builder

or

npm install --save graphql-schema-builder

const graphql = require('graphql');

const {
  buildFields,
  buildTypes,
  getProjection,

  GraphQLJSON
} = require('graphql-schema-builder')(graphql);

function getSchema(resolvers, schema, { customerProvider }) {
  // build types based on existing domain schemas and resolvers
  const types = buildTypes({
    Asset,
    Customer,
    Sensor
  }, resolvers);

  const schemaStore = new Map;

  function buildSubType({ name, fields }) {
    const _name   = `${name}Input`;
    const newType = new GraphQLInputObjectType({ name: _name, fields });
    schemaStore.set(_name, newType);
    return newType;
  }

  function getExistingType(name) {
      return schemaStore.get(`${name}Input`);
  }

  return new GraphQLSchema({
    query: new GraphQLObjectType({
      name: 'RootQueryType',
      fields: {
        customer: {
          type: types.Customer,
          args: {
            id: {
              name: 'id',
              type: new GraphQLNonNull(GraphQLID)
            }
          },

          resolve(_0, { id }, _1, info) {
            const projection = getProjection(info);
            return customerProvider.findById(id, projection);
          }
        },
        customers: {
          type: new GraphQLList(types.Customer),

          resolve(_0, {}, _1, info) {
            const projection = getProjection(info);
            return customerProvider.findAll(projection);
          }
        }
      }
    }),

      mutation: new GraphQLObjectType({
        name: 'Mutation',
        fields: {
          createAsset: {
            type: types.Asset,
            // build arguments based on domain schemas instead of
            // specifying them manually
            args: buildFields(Asset.fields, {
              buildSubType,
              getExistingType
            }),

            resolve(obj, args, source, info) {
              // create asset
            }
          },

          /* more mutations, etc. */
        }
      })
  });
}

Schema example

fields can be either an object or a function accepting { Mixed, ObjectId }. See Mongoose Guide for more details about Schema definition.

const schema = {
  Asset: {
    name:        'Asset',
    description: 'An asset.',

    fields: ({ Mixed, ObjectId }) => ({
      customer: {
        description: 'Customer that this asset belongs to.',

        type:     ObjectId,
        ref:      'Customer',
        required: true
      },

      parent: {
        type:     ObjectId,
        ref:      'Asset',
        required: false
      },

      name: {
        type:     String,
        required: true
      }
    }),

    dynamicFields: ({ ObjectId }) => ({
      sensors: {
        type: [ObjectId],
        ref:  'Sensor'
      }
    })
  },

  Customer: {
    name:        'Customer',
    description: 'A customer.',

    fields: {
      name: {
        description: 'The name of the customer.',

        type:     String,
        required: true
      },

      // Will result in subtype
      metadata: {
        created: {
            type:     Date,
            required: true
        }
      }
    },

    dynamicFields: ({ Mixed, ObjectId }) => ({
      assets: {
        type: [ObjectId],
        ref:  'Asset'
      }
    })
  },

  Sensor: {
    name:        'Sensor',
    description: 'A sensor that must be connected to an asset.',

    fields: ({ Mixed, ObjectId }) => ({
      externalId: {
        type:     String,
        required: false
      },

      asset: {
        description: 'An asset that this sensor is connected to.',

        type:     ObjectId,
        ref:      'Asset',
        required: true
      },

      name: {
        type:     String,
        required: false
      }
    })
  }
}

Resolvers example

A resolver can be either a function or an object. If it's an object it must have a resolve() function and can have args field. args has the same format as a schema. Matching arguments will be passed into the args argument of a resolver.

See also:

• Root Fields & Resolvers
• Passing Arguments

const resolvers = {
  Asset: {
    customer(obj, {}, _, info) {
      const projection = getProjection(info);
      return customerProvider.findById(obj.customer, projection);
    },

    parent(obj, {}, _, info) {
      if (obj.parent != null)
          return null;

      const projection = getProjection(info);
      return assetProvider.findById(obj.parent, projection);
    },

    sensors(obj, {}, _, info) {
      const projection = getProjection(info);
      return sensorProvider.find({ asset: obj.id }, projection);
    },

    // resolver as an object
    measurements: {
      args: {
        resolution: {
          type:     String,
          required: true
        }
      },

      resolve(obj, { resolution }, _, info) {
        const projection = getProjection(info);
        return measurementProvider.find({ resolution }, projection);
      }
    }
  },

  Customer: {
    assets(obj, {}, _, info) {
      const projection = getProjection(info);
      return assetProvider.find({ customer: obj.id }, projection);
    }
  },

  Sensor: {
    asset(obj, {}, _, info) {
      const projection = getProjection(info);
      return assetProvider.findById(obj.parent, projection);
    }
  }
}

Supported schema types

• Array of any of supported types
• Boolean
• Date
• Mixed
• Number
• ObjectId
• String
• Custom GraphQL types passed to buildTypes through getExistingType argument

Examples

See GraphQL example in pubsub-store repository.

License

MIT