Music Projects — Strudel/Hydra/SuperCollider

Organized, reproducible workspace for composing sets from authentic domain docs and cheatsheets. This repo emphasizes privacy-preserving local processing, deterministic outputs, and clear provenance.

Repo Map

• docs/ — Human-facing docs and indexes
  • docs/index.html — Root docs index (entry point)
  • docs/compositions/ — Orchestrated sets, live sets, performance guides
  • docs/cheatsheets/ — Domain cheatsheets generated from processed functions
  • docs/per_domain_faiss_evaluation.md — Dense evaluation notes
• data/raw/ — Scraped/exported authentic inputs (HTML/TXT/JSON)
• data/processed/ — Pipeline outputs
  • chunks/ — Content chunks
  • functions/ — Enhanced function/operator catalogs
  • indexes/ — Retrieval artifacts, harnesses, evaluation reports
• scripts/ — Deterministic pipeline and tooling
  • See scripts/README.md for usage and common commands

Quickstart

1. Create a virtualenv and install deps as needed.
2. Run processing and validation:
   • make process then make validate
3. Build indexes and evaluate retrieval:
   • make clean_index then make evaluate_keyword
4. Generate compositions:
   • make suggest_plan then make orchestrate_set
5. Preview docs:
   • make serve and open http://localhost:8000/docs/index.html

Principles

• No synthetic data; all inputs are authentic public docs/transcripts.
• Reproducible scripts and evaluation; deterministic artifacts saved under data/processed.
• Secure defaults and offline-friendly harnesses (no external CDNs).

Notes

• If markdown links download in your browser, open them in an editor.
• See docs/compositions/index.html for curated Strudel/Tidal/Hydra sets.