Struct BacktestEngine 

pub struct BacktestEngine { /* private fields */ }

Core vectorized engine (MVP).

Takes a (sorted) long-format DataFrame containing at minimum: timestamp, close, signal (bool/f64; value >0 interpreted as desired exposure in units for variable sizing support added in ug9t).

Generalized from unit-size flips (1hr) to exposure-driven for feature/PA sizing parity verification. See run_streaming_simulation for Next path. When BacktestConfig.symbol_col is set, runs independent per-symbol simulations and returns symbol-tagged trades plus per-symbol and portfolio equity curves.

Implementations

impl BacktestEngine

pub fn new(config: BacktestConfig) -> Self

pub fn with_default_costs() -> Self

pub fn backtest_with_report( &self, lf: LazyFrame, ) -> Result<BacktestReport, BacktestError>

Run backtest and attach PerformanceMetrics in a BacktestReport.

pub fn run(&self, lf: LazyFrame) -> Result<BacktestResult, BacktestError>

Run vectorized simulation on a LazyFrame (collected internally for state machine). Input must be sorted ascending by timestamp (then symbol if multi). Returns rich Polars results.

pub fn run_metrics_only( &self, lf: LazyFrame, ) -> Result<PerformanceMetrics, BacktestError>