sngyai/Sequoia-X

Source: https://github.com/sngyai/Sequoia-X

Sequoia-X: The King Returns

A-Share Quantitative Stock Selection System V2

Introduction

Sequoia-X V2 is a quantitative stock selection system for the A-share market, rebuilt from scratch using modern Python engineering standards. The system is built on core design principles: object-oriented architecture, vectorized computation, and incremental data updates. After the market closes each day, it automatically selects stocks and pushes the results to a Feishu (Lark) group.

The data layer uses baostock (http://baostock.com) (free, no registration required, no rate limits) to fetch historical and incremental daily K-line data (backward-adjusted). This data is stored locally in SQLite, completely bypassing web scraping issues from sources like East Money.

Two Operational Modes

python main.py               # Daily Mode: 8-process incremental data update + strategy execution + Feishu push (2-3 minutes)
python main.py --backfill     # Backfill Mode: One-time import of historical K-lines for the entire market (about 12 minutes)

Built-in Strategies

Quick Start

Requirements

• Python >= 3.10

1. Install Dependencies

# Recommended: use uv (fast package manager)
uv sync

# Or use pip
pip install .

2. Configure Environment Variables

cp .env.example .env
# Edit .env and fill in the Feishu Webhook URL

3. First-Time Historical Data Backfill

python main.py --backfill

This takes about 12 minutes to backfill historical backward-adjusted daily K-line data for approximately 5200 A-shares.

4. Daily Operation

python main.py

Recommended to schedule with crontab to run automatically after market close each trading day:

15 19 * * 1-5 cd /root/Sequoia-X && .venv/bin/python main.py >> log.txt 2>&1

Project Structure

Sequoia-X/
├── main.py                      # Entry point: argparse dispatches daily/backfill modes
├── pyproject.toml               # Dependency declaration + ruff/pytest configuration
├── .env.example                 # Environment variable template
├── data/                        # SQLite database (generated at runtime, not in git)
├── sequoia_x/
│   ├── core/
│   │   ├── config.py            # Pydantic-settings configuration management
│   │   └── logger.py            # Rich structured logging
│   ├── data/
│   │   └── engine.py            # Data engine (baostock backfill + incremental sync + SQLite)
│   ├── strategy/
│   │   ├── base.py              # Abstract base class for strategies
│   │   ├── turtle_trade.py      # Turtle Trading strategy
│   │   ├── ma_volume.py         # Moving Average Volume strategy
│   │   ├── high_tight_flag.py   # High Tight Flag strategy
│   │   ├── limit_up_shakeout.py # Limit-Up Shakeout strategy
│   │   ├── uptrend_limit_down.py # Uptrend Limit-Down strategy
│   │   └── rps_breakout.py      # RPS Breakout strategy
│   └── notify/
│       └── feishu.py            # Feishu Webhook push
└── tests/                       # Property-based tests (hypothesis)

Data Notes

• Data Source: baostock (http://baostock.com) (free, no registration required, no rate limits)
• Adjustment Method: Backward-adjusted (hfq) – historical prices remain unchanged, suitable for incremental storage, avoids data discrepancies from corporate actions.
• Storage: Local SQLite (data/sequoia_v2.db), can be directly copied to other machines for use.
• Daily Incremental Updates: 8-process parallel fetching via baostock, completes full market update in 2-3 minutes.

License

MIT