Quant Trading Engine Pro

Problem

Quant analysis setups often separate data, models, and execution: market data requires manual integration, model calls lack stable interfaces, and outputs are hard to route into advice or simulated trading. This skill connects Eastmoney data, local model services, and an output layer so engineers can build an analysis pipeline with MX_APIKEY and REST APIs.

How it works

The data layer uses mx-data, mx-search, and mx-xuangu with one shared MX_APIKEY. The model layer installs Python dependencies with install.py, validates model files, and exposes a local FastAPI service at http://localhost:8000/docs. Core capabilities include:
- Market clustering: K-Means separates bull, bear, and range markets to inform position sizing;
- Prediction and decisioning: /api/v1/analyze and /api/v1/predict combine multiple engine scores;
- Risk and portfolio controls: reliability_screen.py and real_portfolio_manager.py support multi-factor filtering, Kelly sizing, and risk parity;
- Pro features: RL rewards, self-reflection, portfolio advice, and premarket checks are available through /api/v1/pro/* endpoints.
By default, output is analytical advice; with a simulated account, signals can be routed to mx-moni for simulated orders.

Boundaries

The skill does not bundle data sources, so an Eastmoney API key is required. It depends on a Python environment and third-party data, without guarantees of real-time accuracy or completeness. All outputs are for technical research and do not constitute investment advice; users must comply with local regulations and accept their own risk.