DSH HUB
HomePlugin StorePlugin PacksCommunityRankingsResourcesPublish Guide
Plugin source
Back to catalog

vicfei /

vicfei/dsh-nautilus_trader

Topic repository only

NautilusTrader (31.6k★) as DeepSeek Harness tools — professional event-driven backtesting with compiled-core indicators, fee-aware accounting, honest result metadata; live trading guarded off. LGPL-clean posture (unmodified pip dependency). Not affiliated.

★ 0 Stars0 Forks0 IssuesN/A Community rating0 Confirmed installs
View on GitHub
READMESource: main@730d22e2

dsh-Nautilus — a professional backtesting engine as DeepSeek Harness tools

Trend strategies, fees, drawdowns — computed by NautilusTrader's compiled-core indicators and event-driven engine, not hand-rolled arithmetic. Every result carries bars/trades/span/fee metadata; the honest number wins. 专业级回测引擎进 DSH:策略、费率、回撤——出自 NautilusTrader 的编译核心指标 与事件驱动引擎。每个结果带 bars/trades/span/费率元数据,诚实的数字说了算。

DeepSeek Harness License: MIT NautilusTrader LGPL-3.0

Built on NautilusTrader (31.6k★) — independent community bridge, not affiliated. Family: dsh-TradingAgents · dsh-ai-hedge-fund · dsh-Vibe-Trading · dsh-OpenBB · dsh-CCXT.

License posture (upstream is LGPL-3.0 — how we stay clean)

MIT bridge over an unmodified pip dependency: no vendoring, no source modification, user installs nautilus_trader themselves. LGPL explicitly permits this use; the red line (modifying the library without open-sourcing changes) is never crossed. See ROADMAP for details.

Why (Vibe-Trading already has a mini backtest — what's this?)

toy SMA cross in dsh-Vibe-Trading dsh-Nautilus
Engine 20-line hand-rolled loop NautilusTrader compiled-core indicators + event-driven execution
Fees/slippage none ("fee-free structural probe") fee-aware, order-book slippage in M1
Results return + rough DD trades/drawdown/span/fee metadata on every payload
Live trading — guarded off (family fail-closed DNA, M2)

Quick start

pip install --user nautilus_trader
git clone https://github.com/vicfei/dsh-nautilus_trader && cd dsh-Nautilus
python3 scripts/nautilus_bridge.py --run backtest.sma_cross \
  --data /path/to/klines.json --params '{"fast":20,"slow":60,"fee_bps":5}'

Real smoke (2026-11-17, binance BTC/USDT 365d via relay): SMA 20/60 −10.3% (4 trades, max DD −24.8%) vs buy-hold −29.9% — trend-following cut the loss in a bear year but didn't save it. An honest number, honestly labeled.

Status & roadmap

v0 bridge (backtest.sma_cross + indicator.compute) smoke-tested on live data. ROADMAP: M1 full Actor/Strategy wiring + order-book slippage; M2 live trading behind the family's fail-closed guard.

Disclaimer

Not investment advice. Past performance ≠ future results; bar-level execution limits are stated on every payload.

—/ 5

No ratings yet

Manifest verification required

Commit 730d22e2ea03

Community comments

No comments yet. Be the first to write one.

DSH HUB

A community index for DSH plugins. Not an official GitHub or DeepSeek AI product.

CommunityResourcesAPIAbout