Built for Perday CatalogLAB - a live demo of a data story platform for music producers and songwriters.
Advanced music title parsing with version detection for remix, live, and remaster variants. (Tuned for YouTube)
Repo: https://github.com/wmoore012/music_title_parser What it does: Extracts the clean title, version info, and featured artists from messy song titles so CatalogLAB can group songs intelligently.
Song titles from streaming platforms are a mess (Tuned for YouTube):
"Dreams (feat. Snoop Dogg) [Remix]""Dreams - Radio Edit""Dreams (Live at MSG 2024)"
Are these the same song? Kind of. For analytics, you need to know that "Dreams" is the base song and the rest is version metadata.
I built music_title_parser to normalize titles across my catalog. It powers CatalogLAB's song grouping logic and helps identify which versions are getting traction.
- Clean title extraction from any formatting style
- Version detection: remix, live, acoustic, radio edit, remaster, deluxe
- Featured artist parsing with multiple delimiter support (feat., ft., with, x)
- Confidence scoring for ambiguous cases
- Batch processing for catalog-scale operations
pip install music-title-parserOr clone locally:
git clone https://github.com/wmoore012/music_title_parser.git
cd music_title_parser
pip install -e .from music_title_parser import parse_title
result = parse_title("Dreams (feat. Snoop Dogg) [Club Remix]")
print(result.clean_title) # "Dreams"
print(result.featured) # ["Snoop Dogg"]
print(result.version_type) # "remix"
print(result.version_name) # "Club Remix"| Metric | Value |
|---|---|
| Parse speed | 100K titles/sec |
| Accuracy | 98% on streaming datasets |
| Memory | Constant (no caching) |
See BENCHMARKS.md for detailed results.
Built by Wilton Moore for Perday Labs. As a Grammy-nominated producer, I know how many ways the same song can be titled across platforms. This module bridges the gap between messy metadata and clean analytics.
Questions about music metadata or collabs?
- LinkedIn: https://www.linkedin.com/in/wiltonmoore/
- GitHub: https://github.com/wmoore012
MIT License. See LICENSE for details.