OMDB

To aid creating Openmusic-Compatible servers, and for educational purposes.

OMDB (Openmusic Database) is the worlds largest openly downloadable database with metadata for over 154 million songs, 28 million albums, and 5 million artists. (4x larger than MusicBrainz, ~same as Discogs ^[1])

Download the OMDB from OneDrive here.

fulldb.tar 80.3 GB -> 172 GB, Created Oct 2 2024, Expires Dec 21 2024

154,331,239 tracks, 28,046,579 albums, 5,779,932 artists, 154,494,422 playbacks

Download the Lite OMDB (Album.Tracks[0].Views > 1000) from OneDrive here.

litedb.tar 12.8 GB -> 32 GB, Created Oct 2 2024, Expires Dec 21 2024

23,171,125 tracks, 4,773,576 albums, 6,639,289 artists, 23,234,561 playbacks

Copyright Notice

The OMDB is for research purposes only. Public use of the database is permitted, however it is your responsibility to ensure that copyright laws for your country are followed. Failure to do so may result in legal pursuit by the owners of the content you are infringing on. Research your country's copyright laws before proceeding.

How To Use

fulldb.tar is a dump of a PostgreSQL 15 database. To use it, you must first create an empty PostgreSQL 15 database, and then restore using the tar file.

DO NOT unzip the tar file. For that matter, do not unzip/open any file downloaded from the internet, unless you're using the proper tools as shown below.

fulldb.tar restores into a database that is 175 GB, and litedb.tar into 27 GB. Make sure you have sufficient storage before proceeding.

Creating a Database

I recommend using a GUI such as pgAdmin, or you can use the terminal.

This will create an empty database called fulldb, assuming PostgreSQL is installed in /Library:

/Library/PostgreSQL/15/bin/createdb -W fulldb

Restoring via fulldb.tar

Next, write fulldb.tar to the database fulldb:

/Library/PostgreSQL/15/bin/pg_restore -W -d fulldb < fulldb.tar

On my Macbook Pro (M1 Pro 2021), this step takes about 4 hours with this file. Once restored, you can query the database via pgAdmin, or a library such as pg (Node JS) or psycopg2 (Python). Your first step should be to run "ANALYZE" to speed up full-text search.

Database Schema

Artists

• Description: Stores information about artists.
• Columns:
  • ID: TEXT, Primary Key
  • Name: TEXT
  • Profile_Photo: TEXT
  • Subscribers: INT

Albums

• Description: Stores information about albums.
• Columns:
  • ID: TEXT, Primary Key
  • Title: TEXT
  • Artwork: TEXT
  • Type: TEXT
  • Year: INT

Playbacks

• Description: Stores playback data, including the audio and video YouTube ID.
• Columns:
  • ID: SERIAL, Primary Key
  • Audio_YTID: TEXT
  • Video_YTID: TEXT
  • Audio_Live: TEXT
  • Is_Explicit: INT

Tracks

• Description: Stores information about tracks, including indexed search vectors and trigram indexes.
• Columns:
  • ID: TEXT, Primary Key
  • Playback_Clean: INT, References Playbacks(ID)
  • Playback_Explicit: INT, References Playbacks(ID)
  • MXID: TEXT
  • Title: TEXT
  • Runtime_Seconds: INT
  • Views: INT
  • Album_Index: INT
  • Album: TEXT, References Albums(ID)
  • TitleTrgm: TEXT
  • AlbumTrgm: TEXT
  • ArtistsTrgm: TEXT
  • title_vector: tsvector
  • album_vector: tsvector
  • artist_vector: tsvector
  • combined_vector: tsvector

Album_Relation

• Description: Defines an Album's base album. "Album 1" and "Album 1 (Deluxe)" have equal baseIDs.
• Columns:
  • relatedID: TEXT
  • baseID: TEXT

Artist_Album

• Description: Relationship between artists and albums.
• Columns:
  • Artist_ID: TEXT, References Artists(ID)
  • Album_ID: TEXT, References Albums(ID)
  • Composite Primary Key: (Artist_ID, Album_ID)

Artist_Track

• Description: Relationship between artists and tracks.
• Columns:
  • Artist_ID: TEXT, References Artists(ID)
  • Track_ID: TEXT, References Tracks(ID)
  • Composite Primary Key: (Artist_ID, Track_ID)

Features_Track

• Description: Relationship between featured artists and tracks.
• Columns:
  • Artist_ID: TEXT, References Artists(ID)
  • Track_ID: TEXT, References Tracks(ID)
  • Composite Primary Key: (Artist_ID, Track_ID)