Overview

BagelQuant Data is a local research data lake. It stores canonical Parquet files, SQLite metadata, and update history under one root directory.

The public API has three facades:

• lake.admin registers sources and datasets and inspects lake state.
• lake.update plans and runs provider updates.
• lake.query lazily reads stored data.

Datasets are defined only by update behavior: general replaces one dataset file, by_daily updates missing trading dates, and by_asset updates each asset.