波龙股神量化交易系统

🐉 波龙股神量化交易系统 V2.1 — A股/期货一站式量化工作台。多因子选股 · A股回测(T+1/涨跌停) · 华鑫QMT实盘 · 文华财经期货 · 实时风控告警。搜索关键词：波龙股神、量化交易、选股、A股、期货、QMT、文华财经、回测、策略、投资、股票、炒股、量化、量化选股、自动交易、巧未来

Security Scan

This package is a plausible A‑share/futures quant trading toolkit, but there are inconsistencies between its declared metadata and the runtime instructions/config and it asks for trading credentials — review and test carefully before giving real secrets or enabling live trading.

Purpose & Capability

The code and SKILL.md implement a full trading stack (stock picker, backtester, executors for 华鑫QMT and 文华WH8, risk monitor) which legitimately needs Python, market-data tokens (TUSHARE), and optional broker credentials. However the registry summary at the top of the submission lists no required binaries or env vars while the SKILL.md metadata and config files explicitly require python3 and TUSHARE_TOKEN (and optionally QMT/WH8/TB account credentials). That mismatch between declared requirements and the actual instructions/config is an incoherence that should be resolved by the author.

Instruction Scope

The SKILL.md instructs the agent to run local Python scripts, edit local YAML config files, and provide tokens/credentials for data and broker APIs. The instructions reference only trading-relevant files, local service endpoints (defaulting to 127.0.0.1 for broker APIs), and market-data providers (tushare/akshare). There are no instructions to read unrelated system files or exfiltrate data in the provided docs.

Install Mechanism

No install spec is present (instruction-only install), so nothing is downloaded automatically by the skill. Dependencies are typical Python packages installed via pip per the README/SKILL.md (pandas, numpy, tushare, akshare, requests, etc.). This is low-risk from an automated-install perspective, but you should still inspect or pin third-party Python packages before installing.

Credentials

The environment variables referenced in SKILL.md and config (TUSHARE_TOKEN, QMT_ACCOUNT/QMT_PASSWORD, WH8_ACCOUNT/WH8_PASSWORD, PUSH_TOKEN/PUSH_CHAT_ID) are proportional to a trading skill: market-data tokens and broker credentials are expected. The concern is twofold: (1) the registry-level 'Requirements' in the submission initially reported 'none', but the SKILL.md requires TUSHARE_TOKEN and python3 — this discrepancy is confusing; (2) broker credentials are sensitive — give them only to code you trust and prefer paper-mode or separate accounts for testing.

Persistence & Privilege

The skill does not request always:true and does not attempt to modify other skills or systemwide agent settings in the provided files. It runs as-needed scripts and supports simulated (paper) modes in the executors, which is appropriate for a trading toolkit.

What to consider before installing

This repository appears to be a real quant-trading toolkit and thus legitimately needs a market-data token (TUSHARE_TOKEN) and—if you want live trading—broker credentials for QMT/WH8/TB. Two things to check before installing or providing secrets: (1) metadata mismatch — confirm with the author/maintainer which env vars are actually required (the SKILL.md and config expect TUSHARE_TOKEN and optional broker creds, but registry metadata initially showed none); (2) never supply live broker credentials until you have audited the executor code and tested thoroughly in 'paper' (simulated) mode. Recommended steps: run everything in an isolated environment, inspect scripts executor_huaxin.py and executor_wenhua.py for network endpoints and behavior, use a throwaway/test trading account (or only connect to local test endpoints), keep real funds offline until manual review and paper testing succeed, and consider pinning/installing Python dependencies in a virtualenv to avoid supply-chain surprises.

quant-aio · A股期货一站式量化工作台

选股 → 回测 → 实盘 → 风控，一个技能全搞定。

功能概览

┌─────────────────────────────────────────────────────────┐
│                  📈 quant-aio v1.0                      │
├─────────────────────────────────────────────────────────┤
│  📊 选股引擎   多因子筛选 · 行业轮动 · 自选股池管理        │
│  🔁 回测系统   策略编写 · 历史回测 · 绩效报告              │
│  ⚡ 实盘执行   华鑫QMT · 文华财经 · TBQuant3            │
│  🔔 风险监控   仓位监控 · 回撤预警 · 异动推送             │
└─────────────────────────────────────────────────────────┘

环境要求

# Python 3.8+
pip install pandas numpy tushare akshare scipy matplotlib

# tushare pro（推荐，可申请免费token）
pip install tushare

# 华鑫证券QMT Python API
# 安装路径默认: C:\Huaxin\QMT\bin\Lib\site-packages\

# 文华财经WH8 Python API
# 安装路径: 文华财经安装目录下 wh8\python\

# TBQuant3 Python API
# 安装路径: 通联数据官网下载

环境变量

# tushare pro token（必须，申请: https://tushare.pro/register）
export TUSHARE_TOKEN="your_token_here"

# 华鑫QMT（可选，无则跳过实盘模块）
export QMT_ACCOUNT="your_account"
export QMT_PASSWORD="your_password"

# 告警推送（可选）
export PUSH_TOKEN="telegram_bot_token"
export PUSH_CHAT_ID="telegram_chat_id"

模块一：选股引擎

选股策略

# 运行选股（默认：多因子策略）
python scripts/stock_picker.py

# 指定策略
python scripts/stock_picker.py --strategy momentum          # 动量策略
python scripts/stock_picker.py --strategy value             # 价值策略
python scripts/stock_picker.py --strategy growth            # 成长策略
python scripts/stock_picker.py --strategy break_through     # 突破策略

# 自定义参数
python scripts/stock_picker.py --strategy momentum --top 20 --date 2026-03-27

因子配置

编辑 config/factors.yaml 调整因子权重：

factors:
  fundamentals:
    weight: 0.50          # 基本面权重
    pe: [-30, 50]         # PE范围
    roe: [5, 50]          # ROE范围
    revenue_growth: 5      # 营收增速下限%
    profit_growth: 5       # 净利润增速下限%

  technical:
    weight: 0.35          # 技术面权重
    rsi: [30, 70]         # RSI范围
    volume_ratio: 1.5      # 量比下限
    break_20d: true        # 20日均线突破

  sentiment:
    weight: 0.15          # 资金流权重
    north_money: 100      # 北向资金净流入下限（万）
    margin_balance: 100    # 融资余额增速下限%

选股输出

# 输出路径: output/stock_picks_YYYYMMDD.csv

股票代码 | 股票名称 | 综合得分 | PE | ROE | 北向资金 | 入选理由
---------|---------|---------|-----|-----|---------|----------
600519   | 贵州茅台 | 92      | 28  | 32  | +3285万 | 低估值+北向净流入
000858   | 五粮液   | 87      | 22  | 28  | +2156万 | 突破20日均线
...

模块二：回测系统

运行回测

# 完整回测（2015-2026）
python scripts/backtest.py --strategy momentum

# 近期回测（近一年）
python scripts/backtest.py --strategy momentum --period 1y

# 指定股票池
python scripts/backtest.py --strategy momentum --pool 沪深300

# 自定义参数
python scripts/backtest.py --strategy momentum \
  --start 2020-01-01 \
  --end 2026-03-27 \
  --initial_cash 100000 \
  --commission 0.0003 \
  --slippage 0.0001

回测报告

运行后在 output/backtest_report_YYYYMMDD.md 生成报告：

# 📊 回测报告 · 动量策略
- 时间: 2020-01-01 ~ 2026-03-27
- 初始资金: 100,000 元
- 最终净值: 285,600 元
- 总收益率: +185.6%
- 年化收益: 18.3%
- 夏普比率: 1.85
- 最大回撤: -22.4%
- 胜率: 58.3%
- 盈亏比: 1.72
- 总交易次数: 347

# 📈 收益曲线
[图表路径: output/charts/equity_curve.png]

支持指标

类型	指标
趋势	MA/EMA/SMA、MACD、布林带、ADX
动量	RSI、KDJ、CCI、WR
量价	成交量、量比、换手率、OBV
风险	ATR、VaR、Beta

A股特有规则

# 自动处理A股T+1制度
# 自动处理涨跌停（涨停不买入，跌停不卖出）
# 自动处理ST股（可选排除）
# 自动处理停牌（跳过不交易）
# 手续费默认：万3佣金 + 万0.5印花税

模块三：实盘执行

华鑫证券QMT

华鑫证券QMT使用 Python API 直连接口。

from scripts.executor_huaxin import HuaxinExecutor

executor = HuaxinExecutor(
    account="你的账号",
    password="你的密码",
    mode="paper"  # paper=模拟, live=实盘
)

# 下单
executor.buy("600519", price=1850, volume=100)  # 买入贵州茅台
executor.sell("600519", price=1900, volume=100) # 卖出

# 查询
executor.positions()   # 当前持仓
executor.orders()      # 订单状态
executor.balance()     # 账户余额

文华财经WH8

文华财经使用 Python SDK 接口。

from scripts.executor_wenhua import WenhuaExecutor

executor = WenhuaExecutor(
    mode="paper"  # paper=模拟, live=实盘
)

# 期货下单
executor.buy("IF2504", price=3800, lots=1)    # 买入IF股指期货
executor.sell("IC2504", price=5800, lots=1)  # 卖出IC

# 策略执行
executor.run_strategy("momentum_20min")

TBQuant3

TBQuant3通过 Python API 接入。

from scripts.executor_tbquant import TBQuantExecutor

executor = TBQuantExecutor(mode="paper")

# 商品期货
executor.buy("CU2504", price=72000, lots=1)  # 沪铜
executor.sell("RB2504", price=3500, lots=1)  # 螺纹钢

实盘日志

[2026-03-27 09:35:01] 信号: 买入 600519 @ 1850
[2026-03-27 09:35:02] 订单发送成功, OrderID: 12345678
[2026-03-27 09:35:05] 订单成交, 成交价: 1850.2
[2026-03-27 09:35:05] 持仓更新: 600519 x 100股, 成本: 1850.2

模块四：风险监控

监控面板

# 启动实时监控
python scripts/risk_monitor.py

# 启动并推送到Telegram
python scripts/risk_monitor.py --push telegram

风控规则

编辑 config/risk_rules.yaml：

risk_rules:
  # 仓位限制
  max_position_per_stock: 0.20       # 单股最大仓位20%
  max_total_position: 0.80           # 总仓位上限80%
  min_cash_reserve: 0.10             # 最低现金储备10%

  # 回撤限制
  max_daily_loss: 0.02               # 单日最大亏损2%
  max_total_drawdown: 0.15           # 总回撤上限15%
  stop_loss_per_trade: 0.05          # 单笔止损5%

  # 预警规则
  alert_on_position_change: true     # 仓位变化告警
  alert_on_drawdown_threshold: 0.10 # 回撤10%告警
  alert_on_unusual_volume: true       # 异动告警

  # 交易限制
  max_trades_per_day: 10            # 每日最大交易次数
  avoid_limit_up: true              # 避免涨停板买入
  avoid_limit_down: true             # 避免跌停板卖出
  avoid_st_stocks: false            # 排除ST股

风控报告

# 🔔 风控日报 · 2026-03-27

## 持仓状况
| 股票 | 持仓 | 成本 | 现价 | 盈亏 | 仓位 |
|------|------|------|------|------|------|
| 600519 | 500 | 1800 | 1850 | +2.78% | 18.5% |

## 风控指标
- 总仓位: 45.2% ✅ (上限80%)
- 现金储备: 54.8% ✅ (下限10%)
- 今日亏损: -0.8% ✅ (上限2%)
- 总回撤: -3.2% ✅ (上限15%)

## 告警记录
[无]

## 建议
✅ 所有指标正常，继续执行策略

数据源配置

tushare pro（推荐）

# 安装
pip install tushare

# 配置token（在 tushare.pro 注册获取免费token）
# 编辑 config/data_source.yaml

data_source:
  primary: tushare_pro
  tushare:
    token: "${TUSHARE_TOKEN}"  # 设置环境变量
    high_priority: true         # 优先使用日线以上数据

  fallback: akshare
  akshare:
    free: true
    delay: 15                  # 数据延迟15分钟

数据更新

# 更新日线数据
python scripts/update_data.py --freq daily

# 更新分钟线数据
python scripts/update_data.py --freq 5min --days 30

# 全量更新（首次运行）
python scripts/update_data.py --full

目录结构

quant-aio/
├── SKILL.md                      # 本文件
├── README.md                     # 详细使用说明
├── scripts/
│   ├── stock_picker.py          # 选股引擎
│   ├── backtest.py              # 回测系统
│   ├── executor_huaxin.py       # 华鑫QMT执行器
│   ├── executor_wenhua.py       # 文华财经执行器
│   ├── executor_tbquant.py      # TBQuant3执行器
│   ├── risk_monitor.py          # 风险监控
│   ├── update_data.py           # 数据更新
│   └── report_generator.py      # 报告生成
├── config/
│   ├── config.yaml              # 主配置
│   ├── factors.yaml             # 选股因子配置
│   ├── risk_rules.yaml          # 风控规则
│   └── strategies.yaml          # 策略模板
├── references/
│   ├── qmt_api.md              # 华鑫QMT API文档
│   ├── wenhua_api.md           # 文华财经API文档
│   ├── tbquant_api.md          # TBQuant3 API文档
│   ├── tushare_guide.md        # tushare使用指南
│   └── data_sources.md         # 数据源说明
└── output/
    ├── stock_picks/            # 选股结果
    ├── backtest_reports/       # 回测报告
    └── charts/                 # 图表输出

快速开始

# 1. 安装依赖
pip install pandas numpy tushare akshare scipy matplotlib

# 2. 配置（复制并编辑）
cp config/config.yaml.example config/config.yaml

# 3. 设置环境变量
export TUSHARE_TOKEN="你的token"

# 4. 运行选股
python scripts/stock_picker.py --top 20

# 5. 运行回测
python scripts/backtest.py --strategy momentum

# 6. 启动风控监控
python scripts/risk_monitor.py --push telegram

注意事项

⚠️ 实盘交易有风险，请务必先在模拟盘测试稳定后再使用实盘。 ⚠️ 本工具仅供学习和研究使用，盈亏自负。 ⚠️ 请遵守相关法律法规和券商规定。

更新日志

v2.1.1 (2026-04-12)

• 价值投资导向优化：解决追涨杀跌问题
• 增加涨幅过滤：近3月涨幅≤50%，近1月涨幅≤25%
• RPS优化：70-85为最佳区间，排除>90的高位股
• 估值筛选：PE 5-30，PB<5
• 排除高位震荡股和涨停股
• 投资理念：从追涨杀跌转变为价值投资
• 避免选到招商轮船、中远海能等已涨2-3倍的高位股

v2.1.0 (2026-04-11)

• V2.1版本发布

v2.0.0 (2026-03-27)

• 正式命名为"波龙股神量化交易系统"
• 开发者：SimonsTang
• 公司：上海巧未来人工智能科技有限公司
• 开源代码，MIT协议

v1.0.0 (2026-03-27)

• 初始版本
• 支持选股、回测、风控
• 支持华鑫QMT、文华财经WH8、TBQuant3实盘接口
• 支持tushare pro、akshare数据源

---

📜 版权声明

波龙股神量化交易系统 V2.0

• © 2026 SimonsTang / 上海巧未来人工智能科技有限公司 版权所有
• 开源协议：MIT License
• 源码开放，欢迎完善，共同进步
• GitHub：https://github.com/simonstang/bolong-quant
• 联系：[e-mail]
• 网站：https://learn2study.cn

学来学去学习社 · 让学习更有趣更高效