66bc8d12e5
Remove pub state_dim field from DQNConfig and GpuReplayBufferConfig; remove the state_dim field from GpuExperienceCollector. Replace all reads with ml_core::state_layout::STATE_DIM (and STATE_DIM_PADDED for cuBLAS-padded strides). Checkpoint loading now validates saved state_dim against the constant and hard-errors on mismatch. GpuAttentionConfig.state_dim is a distinct attention-feature dim and is left untouched. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
backtesting
Strategy backtesting engine for simulating trading strategies against historical market data.
Key Types
Backtester— main backtesting engineBacktestConfig— simulation configuration (time range, instruments, slippage, commissions)BacktestResults— performance metrics (Sharpe, max drawdown, alpha, beta, Sortino)
Features
- Historical data replay from Parquet files (ticks, order book snapshots, candles)
- Configurable slippage models (fixed, percentage, volume-based)
- Commission modeling (fixed, percentage, per-contract)
- Pluggable strategy interface
Usage
use backtesting::{Backtester, BacktestConfig};
let config = BacktestConfig { /* ... */ };
let results = backtester.run(&strategy)?;