TheMusicTreePipelines
A uv workspace monorepo of Python/Polars pipelines that reconstruct a genre hierarchy from MusicBrainz's flat genre list, using Wikidata's music genre taxonomy as a reference, and export the genre trees and every deduplicated MusicBrainz song for GrowTheMusicTree.
The goal is a reference that classifies all music, so the pipelines keep every song: a song without a genre or without a playable video still ships, so users can link it to one later.
Who it's for
Data engineers and contributors who want to understand or extend how TheMusicTree's genre and song data is built.
Key features
- MusicBrainz pipeline: genre hierarchy from root genre to subgenre to recording
- Wikidata pipeline: genre taxonomy (P279 subclass-of tree) from the public SPARQL endpoint
- Curation pipeline: pulls hand-curated rule lists from the GrowTheMusicTree API
- Gold pipeline: exports genre trees and deduplicated songs, with genre and video when known
- 90% combined test coverage enforced in CI, with Ruff, pytest and pre-commit
Related projects
Feeds GrowTheMusicTree with its genre trees and song catalog.
Technical documentation
Setup, APIs, architecture, and contribution workflows live in each repository's README (and linked resources). This page focuses on the product story and ecosystem fit.