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>
trading_service
Core execution engine for order placement, position keeping, risk integration, and real-time P&L tracking.
Key Types
TradingServiceImpl-- main gRPC serviceOrderExecutor-- high-throughput order handlingPositionManager-- real-time position and P&L trackingRiskIntegration-- pre-trade and post-trade compliance
gRPC Endpoints
PlaceOrder-- submit new ordersCancelOrder-- cancel existing ordersGetPosition-- query current positionsSubscribeMarketData-- market data feedsGetPnlUpdates-- real-time P&L updates
Configuration
GRPC_PORT-- gRPC listen portDATABASE_URL-- PostgreSQL connection stringTLS_ENABLED-- enable TLS 1.3 (default: false)TLS_CERT_PATH,TLS_KEY_PATH,TLS_CA_PATH-- TLS certificatesTLS_REQUIRE_CLIENT_CERT-- mutual TLS (default: false)
Testing
SQLX_OFFLINE=true cargo test -p trading_service --lib