china-stock-quant
# A股量化分析
基于 akshare（免费无需token）的A股量化分析工具包。
## 快速开始
```bash
pip install akshare pandas numpy matplotlib
```
## 工作流
### 1. 获取数据
```python
from scripts.fetch_data import *
# ETF日线
df = fetch_etf_daily("159915", "20250101", "20260301")
# 个股日线
df = fetch_stock_daily("000001", "20250101", "20260301")
# ETF分时（日内做T）
df = fetch_etf_intraday("159915")
# 实时行情
df = fetch_realtime("159915")
```
详见 `references/api-reference.md`
### 2. 计算技术指标
```python
from scripts.technical_indicators import *
# 单指标
df['macd'], df['signal'], df['hist'] = calc_macd(df['close'])
df['k'], df['d'], df['j'] = calc_kdj(df['high'], df['low'], df['close'])
df['rsi'] = calc_rsi(df['close'], period=14)
df['upper'], df['mid'], df['lower'] = calc_bollinger(df['close'])
df['vol_ratio'] = calc_volume_ratio(df['volume'])
# 一键全部
df = add_all_indicators(df)
# 信号检测
signals = detect_signals(df)
```
### 3. 策略回测
```python
from scripts.backtest import *
result = run_backtest(
df,
strategy="grid", # grid / ma_cross / bollinger
initial_capital=100000,
grid_num=10, # 网格数（grid策略）
ma_short=5, ma_long=20, # 均线参数（ma_cross策略）
stop_loss=0.05, # 止损比例
take_profit=0.10, # 止盈比例
)
print(result.summary())
```
### 4. 风险评估
```python
from scripts.backtest import assess_risk
risk = assess_risk(df['close'])
# returns: max_drawdown, sharpe_ratio, annual_volatility, calmar_ratio
```
## 策略库
ETF日内做T策略详解见 `references/strategies.md`，包含：
| 策略 | 适用场景 | 核心逻辑 |
|------|---------|---------|
| 网格交易 | 震荡市 | 价格跌破网格线买入，涨回卖出 |
| 均线交叉 | 趋势市 | 短均线上穿长均线买入，下穿卖出 |
| 布林带回归 | 均值回归 | 触下轨买入，触上轨卖出 |
| 波动率突破 | 突破行情 | ATR放大+价格突破时追入 |
## 风控参数（内置默认值）
```python
RISK_PARAMS = {
"max_position_pct": 0.25, # 单只持仓上限
"stop_loss": 0.05, # 止损线 5%
"take_profit": 0.10, # 止盈线 10%
"max_daily_turnover": 0.05, # 日内做T最大换手
"min_trade_amount": 10000, # 最低交易金额（元）
"max_drawdown_limit": 0.15, # 最大回撤警戒线
}
```
## 资源文件
- `scripts/fetch_data.py` — 数据获取
- `scripts/technical_indicators.py` — 技术指标计算
- `scripts/backtest.py` — 回测引擎+风险评估
- `references/strategies.md` — 策略库详解
- `references/api-reference.md` — akshare接口速查
标签
skill
ai
