Persistra

Persistra is a typed Python toolkit for reproducible, point-in-time financial research. It keeps data acquisition, local storage, feature and factor research, portfolio construction, backtesting, and execution inputs explicit and inspectable.

Start here

1. Install Persistra.
2. Follow the offline quickstart.
3. Create a project when you want a standard local layout.
4. Choose a data source: Alpha Vantage or FRED and ALFRED.

Main workflows

• Acquire and cache data.
• Store and inspect results.
• Build point-in-time datasets and factor models.
• Construct and backtest portfolios.
• Run Monte Carlo research.
• Build and reconcile Trading Engine v1 scenarios.

The examples provide complete offline workflows. The API reference documents every public package.

Scope

Persistra is an offline research library. It does not connect to brokers, place live orders, or infer missing point-in-time assumptions. Trading Engine is a separate deterministic executable; Persistra validates and assembles its current v1 contract but does not implement execution semantics.