### Quick Start

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/README.md

Instructions for cloning the project, installing dependencies, configuring notifications, and running the system.

```bash
# 1. 克隆项目
git clone https://github.com/Dzy-HW-XD/a-share-quant-selector.git
cd a-share-quant-selector

# 2. 安装依赖
pip3 install -r requirements.txt

# 3. 配置钉钉通知（可选）
# 编辑 config/config.yaml 填写 webhook 和 secret

# 4. 首次全量抓取数据
python3 main.py init

# 5. 执行选股（自动更新数据、选股、发送钉钉）
python3 main.py run

# 6. 快速测试（只处理前500只股票）
python3 main.py run --max-stocks 500

# 7. 启动Web界面
python3 main.py web
```

--------------------------------

### Example: Get Strategy Parameters

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/api-reference/web_server.md

Example using curl to fetch the current strategy parameters configuration.

```bash
curl http://localhost:5000/api/strategy-params
```

--------------------------------

### Example: Get Stock List with Parameters

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/api-reference/web_server.md

Example using curl to fetch stock list with specified page and per_page parameters.

```bash
# 获取第1页，每页500只
curl http://localhost:5000/api/stocks?page=1&per_page=500

# 获取第2页，每页100只
curl http://localhost:5000/api/stocks?page=2&per_page=100
```

--------------------------------

### Example: Get K-line Chart

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/api-reference/web_server.md

Example using curl to retrieve a stock's K-line chart as a PNG image.

```bash
# 在浏览器中访问
http://localhost:5000/api/chart/600000

# 或通过curl保存
curl http://localhost:5000/api/chart/600000 -o 600000.png
```

--------------------------------

### Code Example for Module Import

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/README.md

Example code demonstrating module import and usage.

```python
# 注意：示例代码中使用的模块已导入
# 实际使用时需要添加相应的import语句
result = csv_manager.read_stock('600000')
```

--------------------------------

### Example: Execute Stock Selection (All Strategies)

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/api-reference/web_server.md

Example using curl to execute all stock selection strategies with an empty JSON body.

```bash
# 执行所有策略
curl -X POST http://localhost:5000/api/select \
  -H "Content-Type: application/json" \
  -d '{}'
```

--------------------------------

### Example: Execute Stock Selection (Specific Strategy)

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/api-reference/web_server.md

Example using curl to execute a specific stock selection strategy with a maximum number of stocks.

```bash
# 执行特定策略，限制500只股票
curl -X POST http://localhost:5000/api/select \
  -H "Content-Type: application/json" \
  -d '{
    "strategy": "BowlReboundStrategy",
    "max_stocks": 500
  }'
```

--------------------------------

### Workflow Example

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/api-reference/strategy_registry.md

An example demonstrating a typical workflow: initializing the registry and data manager, registering strategies, preparing stock data, running strategies, and processing results.

```python
from strategy.strategy_registry import get_registry
from utils.csv_manager import CSVManager

# 1. Initialize registry and data manager
registry = get_registry("config/strategy_params.yaml")
csv_manager = CSVManager('data')

# 2. Automatically register all strategies
registry.auto_register_from_directory("strategy")

# 3. Prepare stock data
all_stocks = csv_manager.list_all_stocks()
stock_data = {}
for code in all_stocks[:1000]:  # Process first 1000 stocks
    df = csv_manager.read_stock(code)
    if not df.empty:
        stock_data[code] = ('Stock Name', df)

# 4. Execute strategies
results = registry.run_all(stock_data)

# 5. Process results
for strategy_name, signals in results.items():
    print(f"\n{strategy_name}: Selected {len(signals)} stocks")
    for item in signals[:5]:  # Display top 5
        print(f"  {item['code']} {item['name']}")
```

--------------------------------

### Command-line Startup

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/api-reference/web_server.md

Commands to start the web server from the command line.

```bash
# Launch Web Server (default port 5000)
python web_server.py

# Specify port
python web_server.py --port 8080

# Start from main.py
python main.py web

```

--------------------------------

### Correct Configuration Example

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/configuration.md

An example of a correctly configured B1PatternMatch section where weights sum to 1.0.

```yaml
# ✅ 正确
B1PatternMatch:
  weights:
    trend_structure: 0.25
    kdj_state: 0.25
    volume_pattern: 0.25
    price_shape: 0.25   # 总和1.0
```

--------------------------------

### Scenario 3: Quick Testing

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/api-reference/main_entry_point.md

Commands for initializing stocks for testing, running stock selection, and starting a web interface to view results.

```bash
# 初始化100只股票用于测试
python main.py init --max-stocks 100

# 执行选股
python main.py run --max-stocks 100

# 启动Web查看结果
python main.py web
```

--------------------------------

### Example: Update Strategy Parameters

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/api-reference/web_server.md

Example using curl to update strategy parameters with new values.

```bash
curl -X PUT http://localhost:5000/api/strategy-params \
  -H "Content-Type: application/json" \
  -d '{
    "strategy": "BowlReboundStrategy",
    "params": {"N": 3.0, "J_VAL": 25}
  }'
```

--------------------------------

### Combined Usage Examples

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/B1_PATTERN_MATCH.md

Examples of combining B1 matching with other categories or similarity thresholds.

```bash
# Stocks in a falling bowl + B1 match
python3 main.py run --b1-match --category bowl_center

# Stocks near the bull/bear line + high similarity
python3 main.py run --b1-match --category near_duokong --min-similarity 70
```

--------------------------------

### Signal Structure Example

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/README.md

Example of the information structure returned by stock selection.

```python
{
    'date': '2025-11-20',
    'close': 10.85,
    'J': -5.20,
    'category': 'bowl_center',
    'reasons': ['回落碗中']
}
```

--------------------------------

### PatternMatcher Usage Example

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/api-reference/pattern_matcher.md

Example of initializing and using the PatternMatcher, including using default and custom configurations.

```python
from strategy.pattern_matcher import PatternMatcher
from strategy.pattern_config import SIMILARITY_WEIGHTS, MATCH_TOLERANCES

# 使用默认配置（从strategy_params.yaml读取）
matcher = PatternMatcher()

# 使用自定义权重
custom_weights = {
    "trend_structure": 0.25,
    "kdj_state": 0.25,
    "volume_pattern": 0.25,
    "price_shape": 0.25
}
matcher = PatternMatcher(weights=custom_weights)
```

--------------------------------

### Scenario 2: Daily Automation

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/api-reference/main_entry_point.md

Example of scheduling a daily stock selection and DingTalk notification at 15:05.

```bash
# 每天15:05执行，自动选股并发钉钉通知
5 15 * * 1-5 python main.py run >> /var/log/quant.log 2>&1
```

--------------------------------

### Strategy Configuration Example

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/README.md

Example YAML configuration for Bowl Rebound Strategy and B1 Pattern Match.

```yaml
# 碗口反弹策略
BowlReboundStrategy:
  N: 2.4            # 成交量倍数（放量阳线判断）
  M: 20             # 回溯天数（查找关键K线）
  CAP: 4000000000   # 流通市值门槛（40亿）
  J_VAL: 0          # J值上限（KDJ超卖阈值）
  M1: 14            # MA周期1（多空线计算）
  M2: 28            # MA周期2（多空线计算）
  M3: 57            # MA周期3（多空线计算）
  M4: 114           # MA周期4（多空线计算）
  duokong_pct: 3    # 距离多空线百分比（分类用）
  short_pct: 2      # 距离短期趋势线百分比（分类用）

# B1完美图形匹配
B1PatternMatch:
  min_similarity: 60      # 最小相似度阈值（只显示>=此值的股票）
  lookback_days: 25       # 回看天数（匹配时使用的数据天数）
  top_n_results: 15       # 展示Top N个匹配结果（钉钉通知中显示的数量）
  weights:                # 四维相似度权重
    trend_structure: 0.30 # 双线结构
    kdj_state: 0.20       # KDJ状态
    volume_pattern: 0.25  # 量能特征
    price_shape: 0.25     # 价格形态
  tolerances:             # 匹配容差参数
    trend_ratio: 0.10     # 趋势比值容差（±10%）
    price_bias: 10        # 价格偏离容差（±10%）
    trend_spread: 10      # 趋势发散容差（±10%）
    j_value: 30           # J值差异容差（±30）
    drawdown: 15          # 回撤幅度容差（±15%）
```

--------------------------------

### Strategy Parameters - B1PatternMatch - Threshold and Results Example

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/configuration.md

Example of tuning B1PatternMatch parameters for similarity threshold, lookback days, and top results.

```yaml
B1PatternMatch:
  min_similarity: 70     # 只显示>=70%相似的候选
  lookback_days: 30      # 用最近30天的数据匹配
  top_n_results: 10      # 只在钉钉显示Top 10
```

--------------------------------

### RateLimiter Usage Example

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/api-reference/dingtalk_notifier.md

Example of how to instantiate and use the RateLimiter to acquire sending permission.

```python
from utils.dingtalk_notifier import RateLimiter

limiter = RateLimiter(max_per_minute=15, min_interval=3.0)
limiter.acquire()  # 获取发送许可，必要时阻塞等待
```

--------------------------------

### Configuration Error Example: Small M Value

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/configuration.md

An example of an incorrect configuration where the M value in BowlReboundStrategy is too small.

```yaml
# ❌ 错误：M值过小
BowlReboundStrategy:
  M: 3    # M应该>=5
```

--------------------------------

### Configuration Example: Pattern Matching Priority

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/configuration.md

Configuration prioritizing pattern matching.

```yaml
B1PatternMatch:
  min_similarity: 65
  lookback_days: 35
  top_n_results: 20
  weights:
    trend_structure: 0.10
    kdj_state: 0.10
    volume_pattern: 0.20
    price_shape: 0.60   # 形态权重最高
```

--------------------------------

### Configuration Example: Aggressive Stock Selection (Quantity First)

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/configuration.md

Configuration for aggressive stock selection prioritizing quantity.

```yaml
BowlReboundStrategy:
  N: 1.5
  M: 40
  CAP: 2000000000
  J_VAL: 50
  duokong_pct: 3.0
  short_pct: 2.0

B1PatternMatch:
  min_similarity: 50
  weights:
    trend_structure: 0.10
    kdj_state: 0.10
    volume_pattern: 0.30
    price_shape: 0.50
  tolerances:
    trend_ratio: 0.20
    j_value: 50
```

--------------------------------

### PatternFeatures Structure Example

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/README.md

Example of the feature vector structure used for B1 matching.

```python
{
    'trend_structure': {...},
    'kdj_state': {...},
    'volume_pattern': {...},
    'price_shape': [...] 
}
```

--------------------------------

### Redirecting Logs

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/api-reference/main_entry_point.md

Examples of how to redirect program output to files, including saving all output, only errors, or using tee to display and save simultaneously.

```bash
# 保存所有输出到文件
python main.py run >> run.log 2>&1

# 只保存错误
python main.py run 2>> error.log

# 使用tee同时显示和保存
python main.py run 2>&1 | tee run.log
```

--------------------------------

### Configuration Error Example: Weights Not Summing to 1.0

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/configuration.md

An example of an incorrect configuration where the weights in B1PatternMatch do not sum to 1.0.

```yaml
# ❌ 错误：权重不等于1.0
B1PatternMatch:
  weights:
    trend_structure: 0.25
    kdj_state: 0.25
    volume_pattern: 0.25
    price_shape: 0.20   # 总和0.95，不是1.0
```

--------------------------------

### Error Response Example

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/api-reference/web_server.md

Example JSON structure for an error response.

```json
{
    "success": false,
    "error": "Stock code does not exist: 999999"
}

```

--------------------------------

### Match Method Usage Example

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/api-reference/pattern_matcher.md

Example of using the match method to calculate similarity between two stocks' features.

```python
from strategy.pattern_matcher import PatternMatcher
from strategy.pattern_feature_extractor import PatternFeatureExtractor
from utils.csv_manager import CSVManager

csv_manager = CSVManager('data')
extractor = PatternFeatureExtractor(lookback_days=25)
matcher = PatternMatcher()

# 提取候选股特征
candidate_df = csv_manager.read_stock('600000')
candidate_features = extractor.extract(candidate_df)

# 提取案例特征
case_df = csv_manager.read_stock('688799')  # 华纳药厂
case_features = extractor.extract(case_df, lookback_days=25)

# 计算相似度
result = matcher.match(candidate_features, case_features)
print(f"总相似度: {result['total_score']:.2f}%")
for dim, score in result['breakdown'].items():
    print(f"  {dim}: {score:.2f}%")
```

--------------------------------

### 修改选股参数示例

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/README.md

Example of how to modify stock selection parameters in the configuration file.

```yaml
BowlReboundStrategy:
  N: 2.5          # 修改成交量倍数
  J_VAL: 25       # 修改J值阈值
  # ... 其他参数
```

--------------------------------

### Workflow Example: Single Stock Feature Extraction

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/api-reference/pattern_feature_extractor.md

A complete workflow example for extracting features from a single stock.

```python
from strategy.pattern_feature_extractor import PatternFeatureExtractor
from utils.csv_manager import CSVManager

csv_manager = CSVManager('data')
extractor = PatternFeatureExtractor(lookback_days=25)

# Read stock data
df = csv_manager.read_stock('600000')

# Extract features
features = extractor.extract(df)
```

--------------------------------

### 实现自定义策略

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/README.md

Example of implementing custom strategy methods based on BaseStrategy.

```python
def calculate_indicators(self, df) -> pd.DataFrame:
    # 计算指标
    pass

def select_stocks(self, df, stock_name='') -> list:
    # 选股逻辑
    return signals
```

--------------------------------

### Starting the Web Server

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/api-reference/web_server.md

Python code to run the Flask web server.

```python
if __name__ == '__main__':
    app.run(host='0.0.0.0', port=5000, debug=False)

```

--------------------------------

### acquire Method Usage Example

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/api-reference/dingtalk_notifier.md

Example demonstrating how to use the acquire method before sending a DingTalk message and how to handle rate limit errors.

```python
limiter = RateLimiter()

# 发送消息前获取许可
limiter.acquire()
# 发送钉钉消息
result = requests.post(webhook, json=payload)

# 如果遇到限速错误，调用 on_rate_limit_error()
if result.json().get('errcode') == 660026:
    limiter.on_rate_limit_error(retry_count=0)
```

--------------------------------

### Constructor Usage Example

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/api-reference/pattern_feature_extractor.md

Demonstrates how to instantiate the PatternFeatureExtractor with default and custom lookback days.

```python
from strategy.pattern_feature_extractor import PatternFeatureExtractor

# Default 25-day lookback
extractor = PatternFeatureExtractor()

# Custom lookback period
extractor = PatternFeatureExtractor(lookback_days=30)
```

--------------------------------

### Configuration Example: Conservative Stock Selection (Quality First)

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/configuration.md

Configuration for conservative stock selection prioritizing quality.

```yaml
BowlReboundStrategy:
  N: 3.5
  M: 20
  CAP: 5000000000
  J_VAL: 10
  duokong_pct: 0.5
  short_pct: 0.5

B1PatternMatch:
  min_similarity: 75
  weights:
    trend_structure: 0.40
    kdj_state: 0.20
    volume_pattern: 0.20
    price_shape: 0.20
  tolerances:
    trend_ratio: 0.05
    j_value: 15
```

--------------------------------

### Extending a New Strategy

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/README.md

Python code example for creating a new stock selection strategy.

```python
from strategy.base_strategy import BaseStrategy

class MyStrategy(BaseStrategy):
    def __init__(self, params=None):
        super().__init__("我的策略", params)
    
    def calculate_indicators(self, df):
        # 计算指标
        return df
    
    def select_stocks(self, df, stock_name=''):
        # 选股逻辑
        return signals
```

--------------------------------

### Configuration Priority - Command Line Override

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/configuration.md

Example demonstrating how command-line arguments override YAML configuration.

```bash
# 命令行参数覆盖YAML配置
python main.py run --b1-match --min-similarity 75 --lookback-days 40
```

--------------------------------

### Crontab for Scheduled Tasks

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/README.md

Example crontab entry for daily automatic stock selection.

```bash
# 编辑 crontab
crontab -e

# 添加以下行（每天15:05执行，仅工作日）
5 15 * * 1-5 cd /root/quant-csv && /usr/bin/python3 main.py run >> /var/log/quant-csv/run.log 2>&1
```

--------------------------------

### Scenario 4: B1 Perfect Graphic Matching Test

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/api-reference/main_entry_point.md

Examples of using the B1 matching feature for stock selection, including custom parameters.

```bash
# 使用B1匹配执行选股
python main.py run --b1-match --max-stocks 500

# 自定义参数
python main.py run \
  --b1-match \
  --lookback-days 30 \
  --min-similarity 75 \
  --max-stocks 200
```

--------------------------------

### System Information Response

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/api-reference/web_server.md

Example JSON response for the system information API.

```json
{
    "success": true,
    "total_stocks": 5000,
    "stocks_with_data": 4950,
    "latest_update": "2025-11-20",
    "data_dir": "/root/quant-csv/data",
    "strategies": [
        "BowlReboundStrategy"
    ]
}

```

--------------------------------

### init 命令

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/api-reference/main_entry_point.md

命令行接口命令，用于首次全量初始化数据。

```bash
python main.py init [--max-stocks N]
```

--------------------------------

### init 命令 - 使用示例

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/api-reference/main_entry_point.md

演示如何使用init命令进行全量初始化、快速测试或初始化样本数据。

```bash
# 全量初始化
python main.py init

# 快速测试（仅500只）
python main.py init --max-stocks 500

# 样本数据（100只）
python main.py init --max-stocks 100
```

--------------------------------

### API Route: Get Strategy Parameters

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/api-reference/web_server.md

API endpoint to retrieve the current configuration of strategy parameters.

```python
@app.route('/api/strategy-params')
def get_strategy_params()

```

--------------------------------

### Feature Vector Example

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/B1_PATTERN_MATCH.md

Example of a feature vector for Guoxuan High-tech (002074) before a breakout.

```json
{
  "trend_structure": {
    "short_vs_bullbear": 1.0,
    "short_slope": -0.95,
    "price_vs_short_pct": 0.0,
    "price_bias_pct": 0.0,
    "trend_spread_pct": 0.0,
    "is_in_bowl": false
  },
  "kdj_state": {
    "j_value": -11.0,
    "j_position": "低位",
    "k_cross_d": false,
    "j_rebound": true
  },
  "volume_pattern": {
    "avg_volume_ratio": 0.88,
    "volume_trend": "量能平稳",
    "shrink_then_expand": false,
    "max_volume_ratio": 2.1
  },
  "price_shape": {
    "normalized_curve": [0.0, 0.1, 0.05, 0.15, ..., 0.95],
    "max_drawdown": 11.0,
    "breakout_strength": -1.5,
    "overall_trend": "下降"
  }
}
```

--------------------------------

### send_selection_results Usage Example

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/api-reference/dingtalk_notifier.md

Example of how to use the send_selection_results method with prepared data.

```python
notifier = DingTalkNotifier(webhook_url, secret)

# 准备数据
category_signals = {
    'bowl_center': [...],
    'near_duokong': [...],
    'near_short_trend': [...] 
}

# 发送结果
success = notifier.send_selection_results(
    strategy_name="碗口反弹策略",
    category_signals=category_signals,
    stock_names=stock_names,
    stock_data_dict=stock_data_dict
)
```

--------------------------------

### run 命令

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/api-reference/main_entry_point.md

命令行接口命令，执行完整流程：更新数据 → 执行选股 → 发送钉钉通知。

```bash
python main.py run [options]
```

--------------------------------

### DingTalkNotifier Usage Example

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/api-reference/dingtalk_notifier.md

Example of how to instantiate the DingTalkNotifier with webhook URL and secret.

```python
from utils.dingtalk_notifier import DingTalkNotifier

notifier = DingTalkNotifier(
    webhook_url="https://oapi.dingtalk.com/robot/send?access_token=xxx",
    secret="SECxxx"
)
```

--------------------------------

### send_b1_match_results Usage Example

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/api-reference/dingtalk_notifier.md

Example of how to use the send_b1_match_results method with a list of matched results.

```python
matched = [
    {
        'code': '600000',
        'name': '平安银行',
        'similarity': 85.5,
        'matched_case': {'name': '华纳药厂', 'date': '2025-05-12'},
        'breakdown': {...},
        'signal': {...}
    },
    ...
]

success = notifier.send_b1_match_results(
    matched_results=matched,
    stock_names=stock_names,
    stock_data_dict=stock_data_dict,
    top_n=15
)
```

--------------------------------

### run 命令 - 使用示例

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/api-reference/main_entry_point.md

演示如何使用run命令执行完整流程，包括指定股票数量、分类和B1匹配参数。

```bash
# 完整流程
python main.py run

# 快速测试（仅500只股票）
python main.py run --max-stocks 500

# 仅显示回落碗中的股票
python main.py run --category bowl_center

# 启用B1完美图形匹配
python main.py run --b1-match

# 自定义B1参数
python main.py run --b1-match --lookback-days 30 --min-similarity 70
```

--------------------------------

### FAQ: How to run continuously in the background?

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/api-reference/main_entry_point.md

Answer to running continuously in the background using the 'schedule' command with 'nohup'.

```bash
nohup python main.py schedule &
```

--------------------------------

### init_data 方法

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/api-reference/main_entry_point.md

QuantSystem类的方法，用于首次全量初始化所有股票的历史数据。

```python
def init_data(self, max_stocks: int | None = None) -> None
```

--------------------------------

### --version 选项

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/api-reference/main_entry_point.md

命令行接口选项，用于显示程序版本。

```bash
python main.py --version
```

--------------------------------

### init_data 方法 - 使用示例

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/api-reference/main_entry_point.md

演示如何使用init_data方法进行全量初始化或快速测试。

```python
system = QuantSystem()

# 全量初始化（耗时）
system.init_data()

# 快速测试
system.init_data(max_stocks=100)
```

--------------------------------

### QuantSystem 构造函数

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/api-reference/main_entry_point.md

QuantSystem类的构造函数，用于初始化系统并加载配置。

```python
def __init__(self, config_file: str = "config/config.yaml")
```

--------------------------------

### FAQ: How to enable B1 perfect graphic matching?

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/api-reference/main_entry_point.md

Answer to enabling B1 perfect graphic matching using the --b1-match flag.

```bash
python main.py run --b1-match
```

--------------------------------

### _smart_update 方法

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/api-reference/main_entry_point.md

QuantSystem类的方法，用于智能更新数据，仅在市场收盘后执行。

```python
def _smart_update(self, max_stocks: int | None = None, check_latest: bool = True) -> None
```

--------------------------------

### 场景1：第一次使用

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/api-reference/main_entry_point.md

典型使用场景示例，展示首次使用时的初始化和启动Web界面的命令。

```bash
# 1. 初始化数据（全量）
python main.py init

# 2. 启动Web查看
python main.py web
```

--------------------------------

### FAQ: How to select stocks of a specific category?

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/api-reference/main_entry_point.md

Answer to selecting stocks of a specific category using the --category parameter.

```bash
python main.py run --category bowl_center
```

--------------------------------

### Getting a Registered Strategy

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/api-reference/strategy_registry.md

Retrieves a registered strategy by its name.

```python
strategy = registry.get_strategy("BowlReboundStrategy")
if strategy:
    print(f"Found strategy: {strategy.name}")
else:
    print("Strategy not found")
```

--------------------------------

### update 命令

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/api-reference/main_entry_point.md

命令行接口命令，用于增量更新数据。

```bash
python main.py update [--max-stocks N]
```

--------------------------------

### schedule 命令

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/api-reference/main_entry_point.md

命令行接口命令，用于按配置定时执行选股。

```bash
python main.py schedule
```

--------------------------------

### Exception Handling

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/api-reference/pattern_feature_extractor.md

Provides an example of how to handle exceptions during feature extraction.

```python
from strategy.pattern_feature_extractor import PatternFeatureExtractor

extractor = PatternFeatureExtractor()

try:
    features = extractor.extract(df)
    
    if not features.get('trend_structure'):
        print("数据不足，特征提取失败")
    else:
        # 特征成功提取
        pass
        
except Exception as e:
    print(f"特征提取异常: {e}")
```

--------------------------------

### Similarity Threshold Adjustment

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/B1_PATTERN_MATCH.md

Examples of adjusting the minimum similarity threshold for stricter or looser matching.

```bash
# Strict matching (only highly similar)
python3 main.py run --b1-match --min-similarity 75

# Loose matching (show more candidates)
python3 main.py run --b1-match --min-similarity 50
```

--------------------------------

### schedule 命令 - 使用示例

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/api-reference/main_entry_point.md

演示如何使用schedule命令启动定时调度，包括后台运行和crontab配置。

```bash
# 启动定时调度
python main.py schedule

# 后台运行
nohup python main.py schedule > /var/log/quant-scheduler.log 2>&1 &

# 定时任务（crontab）
# 编辑crontab
crontab -e

# 添加行：每天15:05执行（仅工作日）
5 15 * * 1-5 cd /root/quant-csv && /usr/bin/python3 main.py run >> /var/log/quant-csv/run.log 2>&1
```

--------------------------------

### QuantSystem 构造函数 - 使用示例

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/api-reference/main_entry_point.md

演示如何使用QuantSystem类的构造函数，包括使用默认配置和自定义配置。

```python
from main import QuantSystem

# 使用默认配置文件
system = QuantSystem()

# 使用自定义配置文件
system = QuantSystem(config_file="config/custom.yaml")
```

--------------------------------

### YAML Parsing Error Handling

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/errors.md

Example of handling yaml.YAMLError when parsing YAML configuration files.

```python
try:
    with open('config/strategy_params.yaml', 'r') as f:
        config = yaml.safe_load(f)
except yaml.YAMLError as e:
    print(f"配置文件格式错误: {e}")
    config = {}  # 使用默认配置
```

--------------------------------

### Data Type Conversion Error Handling

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/errors.md

Example of handling ValueError during string to float conversion.

```python
try:
    price = float(price_str)
except ValueError:
    print(f"无效的价格值: {price_str}")
    price = None
```

--------------------------------

### Backup and Version Management Script

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/configuration.md

Bash script for creating configuration templates and managing actual configurations.

```bash
# 创建模板
cp config/config.yaml config/config.yaml.template

# 编辑实际配置
cp config/config.yaml.template config/config.yaml
# 填入敏感信息

# .gitignore中排除实际配置
echo "config/config.yaml" >> .gitignore
```

--------------------------------

### Connection Timeout Error Handling

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/errors.md

Example of handling ConnectionTimeout or TimeoutError with a fallback to local cache.

```python
try:
    df = fetcher.get_stock_history('600000')
except (ConnectionError, TimeoutError) as e:
    print(f"网络错误: {e}")
    # 重试或使用本地缓存
    df = csv_manager.read_stock('600000')
```

--------------------------------

### web 命令

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/api-reference/main_entry_point.md

命令行接口命令，用于启动Flask Web服务器。

```bash
python main.py web [--port N] [--host ADDR]
```

--------------------------------

### Extract Method Usage Example

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/api-reference/pattern_feature_extractor.md

Shows how to read stock data and extract features using the PatternFeatureExtractor.

```python
from strategy.pattern_feature_extractor import PatternFeatureExtractor
from utils.csv_manager import CSVManager

csv_manager = CSVManager('data')
df = csv_manager.read_stock('600000')

extractor = PatternFeatureExtractor(lookback_days=25)
features = extractor.extract(df)

print(f"Trend features: {features['trend_structure']}")
print(f"KDJ features: {features['kdj_state']}")
print(f"Volume features: {features['volume_pattern']}")
print(f"Price shape: length {len(features['price_shape'])}")
```

--------------------------------

### Environment Variables

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/api-reference/main_entry_point.md

Demonstrates how to set environment variables for PYTHONPATH and QUANT_DATA_DIR before running the script.

```bash
# 设置Python路径
export PYTHONPATH=/path/to/project:$PYTHONPATH

# 设置数据目录
export QUANT_DATA_DIR=/var/data/stocks

# 运行
python main.py run
```

--------------------------------

### Lookback Days Selection

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/B1_PATTERN_MATCH.md

Examples of adjusting the lookback days for different trading styles (short-term, medium-term, long-term).

```bash
# Short-term consolidation (1-2 weeks)
python3 main.py run --b1-match --lookback-days 15

# Medium-term consolidation (3-4 weeks) - Default
python3 main.py run --b1-match --lookback-days 25

# Long-term consolidation (1-2 months)
python3 main.py run --b1-match --lookback-days 40
```

--------------------------------

### web 命令 - 使用示例

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/api-reference/main_entry_point.md

演示如何使用web命令启动Web服务，包括指定端口和IP地址。

```bash
# 启动（默认5000端口）
python main.py web

# 指定端口
python main.py web --port 8080

# 指定IP和端口
python main.py web --host 192.168.1.100 --port 9000
```

--------------------------------

### update 命令 - 使用示例

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/api-reference/main_entry_point.md

演示如何使用update命令更新所有股票或指定数量的股票。

```bash
# 更新所有股票
python main.py update

# 快速测试
python main.py update --max-stocks 500
```

--------------------------------

### REF Function

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/api-reference/technical_indicators.md

References n periods ago (REF). Gets the value from n days ago.

```python
from utils.technical import REF

prev_close = REF(df['close'], 1)  # 前一日收盘价
volume_ratio = df['volume'] / REF(df['volume'], 1)  # 成交量比率
```

--------------------------------

### Basic Commands

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/README.md

Common commands for initializing, updating, running, and testing the quant selector.

```bash
python3 main.py init
python3 main.py update
python3 main.py run
python3 main.py run --max-stocks 500
python3 main.py run --category bowl_center
python3 main.py web
python3 main.py --version
```

--------------------------------

### Production Environment Deployment

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/api-reference/web_server.md

Commands for deploying the web server in a production environment using Gunicorn or uWSGI.

```bash
# Use gunicorn instead of Flask's built-in server
pip install gunicorn
gunicorn -w 4 -b 0.0.0.0:5000 web_server:app

# Or use uwsgi
pip install uwsgi
uwsgi --http :5000 --wsgi-file web_server.py --callable app --processes 4

```

--------------------------------

### FAQ: How to change the number of stocks executed?

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/api-reference/main_entry_point.md

Answer to changing the number of stocks using the --max-stocks parameter.

```bash
python main.py run --max-stocks 100
```

--------------------------------

### Project Structure

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/README.md

Overview of the project's directory and file structure.

```bash
├── main.py                   # 主程序入口
├── web_server.py             # Web服务器
├── B1_PATTERN_MATCH.md       # B1完美图形匹配详细文档
├── strategy/                 # 策略模块
│   ├── __init__.py
│   ├── base_strategy.py      # 策略基类
│   ├── bowl_rebound.py       # 碗口反弹策略
│   ├── strategy_registry.py  # 策略注册器
│   ├── pattern_config.py     # B1完美图形案例配置
│   ├── pattern_library.py    # B1完美图形库管理
│   ├── pattern_matcher.py    # 相似度计算引擎
│   └── pattern_feature_extractor.py # 特征提取模块
├── utils/               # 工具模块
│   ├── akshare_fetcher.py  # 数据获取
│   ├── csv_manager.py      # CSV数据管理
│   ├── technical.py        # 技术指标(KDJ/EMA/MA等)
│   ├── kline_chart.py      # K线图生成（标准版）
│   ├── kline_chart_fast.py # K线图生成（快速版）
│   └── dingtalk_notifier.py # 钉钉通知
├── config/              # 配置文件
│   ├── config.yaml
│   ├── strategy_params.yaml
│   └── github.yaml
├── web/                 # Web前端
│   ├── templates/
│   └── static/
└── data/                # 股票数据（CSV格式，自动创建）
```

--------------------------------

### DTW Calculation Timeout Handling

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/errors.md

Example of checking for the fastdtw library and falling back to a simplified DTW if not found.

```python
try:
    from fastdtw import fastdtw
    HAS_FASTDTW = True
except ImportError:
    HAS_FASTDTW = False
    print("⚠️ fastdtw 未安装，将使用简化版DTW")
```

--------------------------------

### Feature Extraction Failure Handling

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/errors.md

Example of handling a ValueError when feature extraction fails due to insufficient data.

```python
extractor = PatternFeatureExtractor()
features = extractor.extract(df)

if not features.get('trend_structure'):
    print("特征提取失败，数据不足")
    features = extractor._empty_features()
```

--------------------------------

### StrategyRegistry Constructor

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/api-reference/strategy_registry.md

Instantiates the StrategyRegistry, optionally loading parameters from a specified file.

```python
from strategy.strategy_registry import StrategyRegistry

# Use default configuration file
registry = StrategyRegistry()

# Use custom configuration file
registry = StrategyRegistry("config/custom_params.yaml")
```

--------------------------------

### DingTalk Notification Format - Normal Stock Selection

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/README.md

Example format for stock selection results sent to DingTalk.

```text
🎯 BowlReboundStrategy:
N: 2.4 (成交量倍数)
M: 20 (回溯天数)
CAP: 4000000000 (40亿市值门槛)
J_VAL: 0 (J值上限)
duokong_pct: 3
short_pct: 2
M1: 14 (MA周期)
M2: 28 (MA周期)
M3: 57 (MA周期)
M4: 114 (MA周期)

⏰ 2026-03-02 15:30

🥣 回落碗中: 2 只
📊 靠近多空线: 1 只
📈 靠近短期趋势线: 0 只
📈 共选出: 3 只

---

### 📊 000001 平安银行
**分类**: 🥣 回落碗中
**价格**: 10.85 | **J值**: -7.65
**关键K线日期**: 02-28
**入选理由**: 回落碗中

[K线图图片]
```

--------------------------------

### CSV File Read/Write Error Handling

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/errors.md

Example of handling FileNotFoundError and PermissionError when reading CSV files, with a fallback to network fetch.

```python
try:
    df = csv_manager.read_stock('600000')
except FileNotFoundError:
    print("本地缓存不存在，从网络获取")
    df = fetcher.get_stock_history('600000')
except PermissionError:
    print("无权限访问数据目录")
    # 请求权限或更改目录
```

--------------------------------

### Get Registry Instance

Source: https://github.com/dzy-hw-xd/a-share-quant-selector/blob/main/_autodocs/api-reference/strategy_registry.md

Retrieves the global strategy registry instance using a singleton pattern. It can optionally load parameters from a specified file.

```python
from strategy.strategy_registry import get_registry

# Get the global registry
registry = get_registry()

# Automatically register all strategies
registry.auto_register_from_directory("strategy")

# Getting it again later returns the same instance
registry2 = get_registry()
assert registry is registry2  # True
```