Deprecation Notice

There are many use-cases that are not supported by this library, it is recommended to use something else like graphql-parse-resolve-info

@yonderbox/graphql-fields

Turns GraphQLResolveInfo into a map of the requested fields. Flattens all fragments and duplicated fields into a neat object to easily see which fields were requested at any level. Takes into account any @include or @skip directives, excluding fields/fragments which are @include(if: $false) or @skip(if: $true).

Usage

Schema Type definition

const graphqlFields = require('@yonderbox/graphql-fields');
const graphql = require('graphql')

const UserType = new graphql.GraphQLObjectType({
    name: 'User',
    fields: {
        profile: {type: new graphql.GraphQLObjectType({
          name: 'Profile',
          fields: {
            firstName: {type: graphql.GraphQLString},
            lastName: {type: graphql.GraphQLString},
            middleName: {type: graphql.GraphQLString},
            nickName: {type: graphql.GraphQLString},
            maidenName: {type: graphql.GraphQLString}
          }
        }),
        email: {type: graphql.GraphQLString},
        id: {type: graphql.GraphQLID}
    }
});

module.exports = new GraphQLSchema({
    query: new GraphQLObjectType({
        name: 'Query',
        fields: () =>
            Object.assign({
                user: {
                    type: UserType,
                    resolve(root, args, context, info) {
                        console.log(
                            JSON.stringify(graphqlFields(info), null, 2)
                        );
                        ...
                    }
                }
            })
    })
})

Query

{
  user {
    ...A
    profile {
      ...B
      firstName
    }
  }
}

fragment A on User {
  ...C
  id,
  profile {
    lastName
  }
}

Fragment B on Profile {
  firstName
  nickName @skip(if: true)
}

Fragment C on User {
  email,
  profile {
    middleName
    maidenName @include(if: false)
  }
}

will log

{
  "profile": {
    "firstName": {},
    "lastName": {},
    "middleName": {}
  },
  "email": {},
  "id": {}
}

subfields arguments

To enable subfields arguments parsing, you'll have to provide an option object to the function. This feature is disable by default.

const graphqlFields = require('@yonderbox/graphql-fields');
const fieldsWithSubFieldsArgs = graphqlFields(info, {}, { processArguments: true });

For each subfield w/ arguments, a __arguments property will be created. It will be an array with the following format:

[
    {
        arg1Name: {
            kind: ARG1_KIND,
            value: ARG1_VALUE,
        },
    },
    {
        arg2Name: {
            kind: ARG2_KIND,
            value: ARG2_VALUE,
        }
    }
]

The kind property is here to help differentiate value cast to strings by javascript clients, such as enum values.

Exclude specific fields

Most of the time we don't need __typename to be sent to backend/rest api, we can exclude __typename using this:

const graphqlFields = require('@yonderbox/graphql-fields');
const fieldsWithoutTypeName = graphqlFields(info, {}, { excludedFields: ['__typename'] });

Include alias name

This option returns alias with field name:

const graphqlFields = require('@yonderbox/graphql-fields');
const fieldsWithoutTypeName = graphqlFields(info, {}, { includeAliasName: true });

Request:

{
  person {
    aliasName: name
  }
}

Parsed info object:

{
  aliasName:name: {}
}

Why

An underlying REST api may only return fields based on query params.

{
  user {
    profile {
      firstName
    },
    id
  }
}

should request /api/user?fields=profile,id

while

{
  user {
    email
  }
}

should request /api/user?fields=email

Implement your resolve method like so:

resolve(root, args, context, info) {
    const topLevelFields = Object.keys(graphqlFields(info));
    return fetch(`/api/user?fields=${topLevelFields.join(',')}`);
}

Tests

npm test