2efedcd6b8
Five test files migrated to clear the last cargo check --all-targets errors: - tests/multi_horizon_loader.rs:22,64 — hardcoded [usize;5] horizons literal → ml_alpha::heads::HORIZONS; for-loop bounds 0..5 → 0..N_HORIZONS - tests/output_smoothness_grad_finite_diff.rs:198 — [0.1, 0.3, 1.0, 3.0, 10.0] → [0.1, 1.0, 10.0] preserving 100× span across horizons - src/data/loader.rs:560,640 (inline lib tests) — hardcoded [30,100,300, 1000,6000] literal → crate::heads::HORIZONS - tests/perception_overfit.rs:318,358 (audit-discovered 5-isms) — cfg.seq_len * 5 → cfg.seq_len * N_HORIZONS - tests/gpu_log_ring_invariants.rs:173 (audit-discovered) — payload field name v["payload"]["raw_h30"] → "raw_h10" (matches gpu_log.rs schema migrated in Task 5) cargo check --workspace --all-targets: clean (only pre-existing third-party cudarc cupti example error, unrelated). cargo test -p ml-alpha --lib: 33 passed, 0 failed, 6 ignored — baseline. Golden fixtures deferred to runtime regeneration: - tests/fixtures/perception_forward_golden.bin (644 → 388 bytes post-rebase). Test is #[ignore]-d and rewrites if missing; regenerate during Task 9 local validation by deleting the .bin and re-running with --ignored. gpu_log.rs migration verified complete by Task 5 (no remnant 5-horizon field names in payload_json decoders for RT_INPUT/RT_STATE/RT_OUTPUT). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Foxhunt
Production HFT trading system in Rust.
Architecture
The workspace contains 32 crates organized as follows:
Core Libraries (16)
| Crate | Purpose |
|---|---|
trading_engine |
Order processing, FIX 4.4, IB TWS, SIMD, RDTSC timing |
risk |
VaR, Kelly, circuit breakers, kill switches, compliance |
risk-data |
Risk data types and shared structures |
trading-data |
Trading data types |
ml |
DQN Rainbow, PPO, TFT, Mamba2, ensemble inference |
ml-data |
ML data types and feature definitions |
data |
Market data ingestion and storage |
backtesting |
Replay engine, strategy tester |
adaptive-strategy |
Ensemble execution, microstructure analysis |
common |
Shared types, resilience, error handling |
storage |
S3 and local model storage |
model_loader |
Model serialization and loading |
market-data |
Market data feed handlers |
database |
PostgreSQL access layer (SQLx) |
config |
Configuration management |
tli |
CLI commands and tooling |
Services (8)
| Service | Purpose |
|---|---|
backtesting_service |
gRPC backtesting service |
broker_gateway_service |
FIX routing, broker connectivity |
trading_service |
Core trading operations |
ml_training_service |
Model training orchestration |
data_acquisition_service |
Market data acquisition |
trading_agent_service |
Autonomous trading agents |
api_gateway |
gRPC API gateway with auth |
web-gateway |
Axum REST + WebSocket gateway |
Frontend
web-dashboard/ -- React 19 + TypeScript + Vite + TradingView charts.
Building
# Check compilation (no PostgreSQL required)
SQLX_OFFLINE=true cargo check --workspace
# Run tests for a specific crate
SQLX_OFFLINE=true cargo test -p <crate> --lib
# Clippy
SQLX_OFFLINE=true cargo clippy --workspace
ML Models
Four production model architectures on Candle v0.9.1 with CUDA:
- DQN Rainbow -- Deep Q-Network with prioritized replay, dueling heads, noisy nets
- PPO -- Proximal Policy Optimization with GAE and LSTM policies
- TFT -- Temporal Fusion Transformer for multi-horizon forecasting
- Mamba2 -- State space model for sequence prediction
Each model has a standalone trainer and a UnifiedTrainable adapter for the hyperopt pipeline.
Infrastructure
- Git: Gitea at
git.fxhnt.ai(Tailscale-only), Scaleway DEV1-S - Observability: OpenTelemetry OTLP (env
OTEL_EXPORTER_OTLP_ENDPOINT) - Database: PostgreSQL with SQLx offline mode for CI
License
Proprietary. All rights reserved.