Taste Graph visualizes a listener's Spotify history as clustered artist neighborhoods.
The backend now includes a FastAPI-oriented graph-generation pipeline under backend/. It is designed to:
- fetch Spotify profile, recent listening, top artists, and top tracks through
services/spotify.py - train or synthesize artist embeddings in
services/embeddings.py - reduce and cluster artists in
services/clustering.py - label clusters in
services/labeling.py - cache serialized graph JSON through
db/supabase.py
Install the Python dependencies from backend/requirements.txt, configure the env vars listed in PLAN.md, and serve backend.main:app with your ASGI runner of choice.