[Roadmap] Dynamic universe and point-in-time stock screening

[brokermr810]

Problem

Strategies can consume saved universes, but contributors and users need a first-class way to define, refresh, and reproduce dynamic stock universes using rules such as price, market capitalization, liquidity, volatility, and listing age.

Desired outcome

Create a point-in-time universe pipeline shared by research, backtests, and live strategy runs. A historical run must only see constituents and inputs that were available at that timestamp.

Initial scope

• Define a versioned universe rule contract.
• Support composable AND/OR filters for an initial set of stable market fields.
• Add daily, weekly, monthly, and explicit N-day refresh policies.
• Persist dated constituent snapshots and rule metadata.
• Expose the same saved universe to Strategy API V2, backtests, and live runtimes.
• Record data source, timestamp, and exclusion reasons for auditability.

Acceptance criteria

• An RFC documents the rule schema, snapshot model, and look-ahead protections.
• A deterministic fixture produces the same constituents in research and backtest paths.
• Historical membership changes are covered by tests.
• Missing market data and provider limits have explicit behavior.
• API changes update OpenAPI and compatibility checks.
• User and contributor documentation is included.

Out of scope

• A promise of complete global fundamentals coverage.
• Runtime installation of arbitrary data-provider libraries.
• A visual no-code builder in the first backend contribution.

Contributing

This is an epic. Do not implement the whole initiative in one pull request. Comment with the specific contract, service, test fixture, or documentation task you want to own and include a short plan. A maintainer will confirm scope before assignment.