Add lightweight screener API for batch symbol evaluation

[TDamiao]

Goal

Provide a lightweight market screener API for batch symbol evaluation without turning the library into a platform.

Why

Users need an operational path to evaluate many symbols/rules in one call using existing indicators/strategies.

Implementation prompt

Add a simple API such as:

• screen(symbols, ruleFn, dataProvider, options)
• evaluateBatch(dataset, ruleFn)

Requirements:

• Async-friendly data provider contract
• Deterministic rule evaluation order
• Structured results per symbol (passed, score?, reason?)

Scope

• Library-side utility only (no UI, no scheduler, no orchestration service)
• Works with indicator and strategy outputs
• Optional concurrency controls and timeout policies

Acceptance criteria

• Users can run one rule across multiple symbols with a single function call
• Output schema is typed and stable
• README includes one screener example
• Tests cover provider failures, empty datasets, and deterministic ordering

Test plan

• Unit tests with mocked provider/rule functions
• Integration smoke test using synthetic OHLCV sets
• Performance sanity check for 100+ symbol batch

[TDamiao]

Tracked by v0.4 master gate: #21.