GraphBrainz

A GraphQL schema, Express server, and middleware for querying the MusicBrainz API. It features an extensible schema to add integration with Discogs, Spotify, Last.fm, fanart.tv, and more!

Try out the live demo! :bulb: Use the “Docs” sidebar, the schema, or the types docs to help construct your query.

Install

Install with npm:

npm install graphbrainz --save

Install with Yarn:

yarn add graphbrainz

GraphBrainz is written and distributed as native ECMAScript modules (ESM) and requires a compatible version of Node.js

Contents

• Usage
  • As a standalone server
  • As middleware
  • As a client
  • Environment Variables
  • Debugging
• Example Queries
  • Pagination
• Questions
• Schema
  • Extending the schema

Usage

This package can be used both as a standalone GraphQL server and as Express middleware supplying a GraphQL endpoint.

As a standalone server

Run the included graphbrainz executable to start the server. The server is configured using environment variables.

$ graphbrainz
Listening on port 3000.

Development mode features like JSON pretty printing and the GraphiQL interface will be enabled unless the server is run with NODE_ENV=production.

Note that if you are not running the standalone server within another Node project, you may wish to install this package globally so that the graphbrainz script is globally available:

npm install -g graphbrainz

As middleware

If you have an existing Express server and want to add this GraphQL service as an endpoint, or you just want more customization, use the middleware.

import express from 'express';
import { middleware as graphbrainz } from 'graphbrainz';

const app = express();

// Use the default options:
app.use('/graphbrainz', graphbrainz());

// or, pass some options:
app.use('/graphbrainz', graphbrainz({
  client: new MusicBrainz({ ... }),
  graphiql: true,
  ...
}));

app.listen(3000);

The graphbrainz middleware function accepts the following options:

• client: A custom API client instance to use. See the client submodule for help with creating a custom instance. You probably only need to do this if you want to adjust the rate limit and retry behavior.
• Any remaining options are passed along to the standard GraphQL middleware. See the express-graphql documentation for more information.

As a client

If you just want to make queries from your app without running a separate server or exposing a GraphQL endpoint, use the GraphBrainz schema with a library like GraphQL.js. You just need to create the context object that the GraphBrainz resolvers expect, like so:

import { graphql } from 'graphql';
import { MusicBrainz, createContext, baseSchema } from 'graphbrainz';

const client = new MusicBrainz();
const context = createContext({ client });

graphql(
  schema,
  `
    {
      lookup {
        releaseGroup(mbid: "99599db8-0e36-4a93-b0e8-350e9d7502a9") {
          title
        }
      }
    }
  `,
  null,
  context
)
  .then((result) => {
    const { releaseGroup } = result.data.lookup;
    console.log(`The album title is “${releaseGroup.title}”.`);
  })
  .catch((err) => {
    console.error(err);
  });

Environment Variables

• MUSICBRAINZ_BASE_URL: The base MusicBrainz API URL to use. Change this if you are running your own MusicBrainz mirror. Defaults to http://musicbrainz.org/ws/2/.
• GRAPHBRAINZ_PATH: The URL route at which to expose the GraphQL endpoint, if running the standalone server. Defaults to /.
• GRAPHBRAINZ_CORS_ORIGIN: The value of the origin option to pass to the CORS middleware. Valid values are true to reflect the request origin, a specific origin string to allow, * to allow all origins, and false to disable CORS (the default).
• GRAPHBRAINZ_CACHE_SIZE: The maximum number of REST API responses to cache. Increasing the cache size and TTL will greatly lower query execution time for complex queries involving frequently accessed entities. Defaults to 8192.
• GRAPHBRAINZ_CACHE_TTL: The maximum age of REST API responses in the cache, in milliseconds. Responses older than this will be disposed of (and re-requested) the next time they are accessed. Defaults to 86400000 (one day).
• GRAPHBRAINZ_GRAPHIQL: Set this to true if you want to force the GraphiQL interface to be available even in production mode.
• GRAPHBRAINZ_EXTENSIONS: A JSON array of module paths to load as extensions.
• PORT: Port number to use, if running the standalone server.

When running the standalone server, dotenv is used to load these variables from a .env file, if one exists in the current working directory. This just makes it more convenient to launch the server with certain settings. See the dotenv package for more information.

Debugging

The DEBUG environment variable can be used to enable logging for all (or just some) of this package’s submodules:

$ DEBUG=graphbrainz:* graphbrainz

See the debug package for more information.

Example Queries

Nirvana albums and each album’s singles (try it):

query NirvanaAlbumSingles {
  lookup {
    artist(mbid: "5b11f4ce-a62d-471e-81fc-a69a8278c7da") {
      name
      releaseGroups(type: ALBUM) {
        edges {
          node {
            title
            firstReleaseDate
            relationships {
              releaseGroups(type: "single from") {
                edges {
                  node {
                    target {
                      ... on ReleaseGroup {
                        title
                        firstReleaseDate
                      }
                    }
                  }
                }
              }
            }
          }
        }
      }
    }
  }
}

Pagination

The first five labels with “Apple” in the name (try it):

query AppleLabels {
  search {
    labels(query: "Apple", first: 5) {
      ...labelResults
    }
  }
}

fragment labelResults on LabelConnection {
  pageInfo {
    endCursor
  }
  edges {
    cursor
    node {
      mbid
      name
      type
      area {
        name
      }
    }
  }
}

…and the next five, using the endCursor from the previous result (try it):

query AppleLabels {
  search {
    labels(query: "Apple", first: 5, after: "YXJyYXljb25uZWN0aW9uOjQ=") {
      ...labelResults
    }
  }
}

Who the members of the band on an Apple Records release married, and when (try it):

query AppleRecordsMarriages {
  search {
    labels(query: "Apple Records", first: 1) {
      edges {
        node {
          name
          disambiguation
          country
          releases(first: 1) {
            edges {
              node {
                title
                date
                artists {
                  edges {
                    node {
                      name
                      ...bandMembers
                    }
                  }
                }
              }
            }
          }
        }
      }
    }
  }
}

fragment bandMembers on Artist {
  relationships {
    artists(direction: "backward", type: "member of band") {
      edges {
        node {
          type
          target {
            ... on Artist {
              name
              ...marriages
            }
          }
        }
      }
    }
  }
}

fragment marriages on Artist {
  relationships {
    artists(type: "married") {
      edges {
        node {
          type
          direction
          begin
          end
          target {
            ... on Artist {
              name
            }
          }
        }
      }
    }
  }
}

Images of Tom Petty provided by various extensions (try it):

query TomPettyImages {
  lookup {
    artist(mbid: "5ca3f318-d028-4151-ac73-78e2b2d6cdcc") {
      name
      mediaWikiImages {
        url
        objectName
        descriptionHTML
        licenseShortName
      }
      fanArt {
        thumbnails {
          url
          likeCount
        }
      }
      theAudioDB {
        logo
        biography
      }
    }
  }
}

You can find more example queries in the schema tests.

Questions

What’s with the cumbersome edges/node nesting? Why first/after instead of limit/offset? Why mbid instead of id?

You can thank Relay for that; these are properties of a Relay-compliant schema. The schema was originally designed to be more user-friendly, but in the end I decided that being compatible with Relay was a worthwhile feature. I agree, it’s ugly.

The GraphBrainz schema includes an extra nodes field on every connection type. If you only want the nodes and no other fields on edges, you can use nodes as a shortcut.

Don’t forget that you can also use GraphQL aliases to rename fields to your liking. For example, the following query renames edges, node, and mbid to results, releaseGroup, and id, respectively:

query ChristmasAlbums {
  search {
    releaseGroups(query: "Christmas") {
      results: edges {
        releaseGroup: node {
          id: mbid
          title
        }
      }
    }
  }
}

Why does my query take so long?

It’s likely that your query requires multiple round trips to the MusicBrainz REST API, which is subject to rate limiting. While the query resolver tries very hard to fetch only the data necessary, and with the smallest number of API requests, it is not 100% optimal (yet). Make sure you are only requesting the fields you need and a reasonable level of nested entities – unless you are willing to wait.

You can also set up a local MusicBrainz mirror and configure GraphBrainz to use that with no rate limiting.

Schema

The types document is the easiest to browse representation of the schema, or you can read the schema in GraphQL syntax.

Extending the schema

The GraphBrainz schema can easily be extended to add integrations with third-party services. See the Extensions docs for more info.