Vnpy Export

• You have completed a Vibe-Trading backtest run and want to export the strategy as a .py file loadable in the vnpy CTA Strategy App.
• You want to run a strategy in vnpy's CTA backtesting engine using BacktestingEngine with proper parameter and variable declarations.
• You are trading A-share equities (*.SZSE / *.SSE), Chinese futures (*.CFFEX, *.SHFE, etc.), or crypto (*.BINANCE) and need the correct vnpy conventions for each asset class.
• You have a strategy description (no existing backtest run) and want the skill to generate a compliant CtaTemplate subclass from scratch.

• You are targeting a non-CTA vnpy app (e.g., vnpy's Spread Trading or Option Master apps) — this skill only generates CtaTemplate subclasses.
• Your strategy relies on order-book or Level-2 tick logic that cannot be represented in an on_bar / on_tick flow without significant manual rework.

Reads config.json and code/signal_engine.py from an existing Vibe-Trading run, then translates the full signal logic into a CtaTemplate subclass saved to artifacts/vnpy_strategy/<StrategyName>Strategy.py.

When no backtest run exists, generates a compliant CtaTemplate class directly from a plain-language strategy description and writes it to the same output path.

Automatically applies the correct vt_symbol format, position units, and order direction rules for A-share stocks (buy/sell only), futures (all four directions), and crypto — no manual adjustment needed.

Maps common pandas and TA-Lib indicator calls (rolling mean, EWM, RSI, MACD, Bollinger Bands, ATR, Donchian, and more) to their ArrayManager equivalents, avoiding look-ahead bias and improving runtime performance.

Handles strategies that combine multiple timeframes (e.g., daily trend filter + intraday entry) by wiring multiple BarGenerator and ArrayManager instances with the correct callback chain.

Before saving, validates the output against a built-in checklist: correct class naming, matching parameter/variable declarations, cancel_all() and put_event() placement, early-return on not am.inited, and adequate load_bar warmup depth.

A quant has finished backtesting a momentum strategy in Vibe-Trading and wants to run it live in vnpy. The skill reads the run's config and signal engine, then emits a ready-to-load CtaTemplate .py file with all parameters and indicators correctly wired.

A user describes a dual-MA crossover strategy for CSI 300 futures in plain language. The skill writes a complete, compliant CtaTemplate subclass — including BarGenerator, ArrayManager indicator calls, and proper long/short order logic — without requiring an existing codebase.

After export, the generated file can be dropped into a vnpy project and run immediately with BacktestingEngine, using the provided boilerplate for setting commission, slippage, capital, and date range.

A strategy that uses a daily trend filter combined with 5-minute entries is exported with correctly structured nested BarGenerator callbacks and separate ArrayManager instances for each timeframe.