commit;
This commit is contained in:
parent
b3332eb920
commit
a4ef9df315
|
|
@ -0,0 +1,831 @@
|
|||
#!/usr/bin/env python3
|
||||
# -*- coding: utf-8 -*-
|
||||
|
||||
"""
|
||||
科技主题基本面因子选股策略--这里就是入口--请执行这个文件！
|
||||
整合企业生命周期、财务指标和平均距离因子分析
|
||||
"""
|
||||
|
||||
import sys
|
||||
import pandas as pd
|
||||
import numpy as np
|
||||
import logging
|
||||
from typing import Dict, List, Tuple
|
||||
from pathlib import Path
|
||||
from sqlalchemy import create_engine, text
|
||||
from datetime import datetime, timedelta
|
||||
import math
|
||||
|
||||
# 添加项目根路径到Python路径
|
||||
project_root = Path(__file__).parent.parent.parent
|
||||
sys.path.append(str(project_root))
|
||||
|
||||
# 导入依赖的工具类
|
||||
from src.quantitative_analysis.company_lifecycle_factor import CompanyLifecycleFactor
|
||||
from src.quantitative_analysis.financial_indicator_analyzer import FinancialIndicatorAnalyzer
|
||||
from src.quantitative_analysis.average_distance_factor import AverageDistanceFactor
|
||||
# from src.valuation_analysis.config import DB_URL
|
||||
|
||||
|
||||
class DateHelper:
|
||||
"""日期处理工具类，根据输入日期计算应该使用的财务数据日期"""
|
||||
|
||||
@staticmethod
|
||||
def get_annual_date(target_date: str) -> str:
|
||||
"""
|
||||
获取应该使用的年度数据日期
|
||||
|
||||
规则：5月1日之前用上上年数据，5月1日之后用上年数据
|
||||
例如：2025-11-03 -> 2024-12-31 (上年年报)
|
||||
2025-04-30 -> 2023-12-31 (上上年年报)
|
||||
|
||||
Args:
|
||||
target_date: 目标日期，格式 YYYY-MM-DD
|
||||
|
||||
Returns:
|
||||
年度数据日期，格式 YYYY-MM-DD
|
||||
"""
|
||||
date_obj = datetime.strptime(target_date, '%Y-%m-%d')
|
||||
year = date_obj.year
|
||||
month = date_obj.month
|
||||
day = date_obj.day
|
||||
|
||||
# 5月1日之前用上上年数据
|
||||
if month < 5 or (month == 5 and day < 1):
|
||||
annual_year = year - 2
|
||||
else:
|
||||
annual_year = year - 1
|
||||
|
||||
return f"{annual_year}-12-31"
|
||||
|
||||
@staticmethod
|
||||
def get_quarter_date(target_date: str) -> str:
|
||||
"""
|
||||
获取应该使用的季度数据日期
|
||||
|
||||
规则：
|
||||
- 5月1日发去年的年报和今年一季报 -> 使用一季报 (YYYY-03-31)
|
||||
- 8月31日发半年报 -> 使用半年报 (YYYY-06-30)
|
||||
- 10月31日发三季报 -> 使用三季报 (YYYY-09-30)
|
||||
- 其他时间使用最新可用季度
|
||||
|
||||
Args:
|
||||
target_date: 目标日期，格式 YYYY-MM-DD
|
||||
|
||||
Returns:
|
||||
季度数据日期，格式 YYYY-MM-DD
|
||||
"""
|
||||
date_obj = datetime.strptime(target_date, '%Y-%m-%d')
|
||||
year = date_obj.year
|
||||
month = date_obj.month
|
||||
day = date_obj.day
|
||||
|
||||
# 5月1日之前：使用上上年的年报或上年的三季报
|
||||
if month < 5 or (month == 5 and day < 1):
|
||||
# 使用上年的三季报
|
||||
return f"{year - 1}-09-30"
|
||||
# 5月1日到8月31日：使用今年一季报
|
||||
elif month < 8 or (month == 8 and day <= 31):
|
||||
return f"{year}-03-31"
|
||||
# 8月31日到10月31日：使用半年报
|
||||
elif month < 10 or (month == 10 and day <= 31):
|
||||
return f"{year}-06-30"
|
||||
# 10月31日之后：使用三季报
|
||||
else:
|
||||
return f"{year}-09-30"
|
||||
|
||||
@staticmethod
|
||||
def get_lifecycle_year(target_date: str) -> int:
|
||||
"""
|
||||
获取生命周期分析应该使用的年份
|
||||
|
||||
Args:
|
||||
target_date: 目标日期，格式 YYYY-MM-DD
|
||||
|
||||
Returns:
|
||||
分析年份
|
||||
"""
|
||||
date_obj = datetime.strptime(target_date, '%Y-%m-%d')
|
||||
year = date_obj.year
|
||||
month = date_obj.month
|
||||
day = date_obj.day
|
||||
|
||||
# 5月1日之前用上上年，之后用上年
|
||||
if month < 5 or (month == 5 and day < 1):
|
||||
return year - 2
|
||||
else:
|
||||
return year - 1
|
||||
|
||||
# 设置日志
|
||||
logging.basicConfig(
|
||||
level=logging.INFO,
|
||||
format='%(asctime)s - %(name)s - %(levelname)s - %(message)s'
|
||||
)
|
||||
logger = logging.getLogger(__name__)
|
||||
|
||||
|
||||
class TechFundamentalFactorStrategy:
|
||||
"""科技主题基本面因子选股策略"""
|
||||
|
||||
def __init__(self, target_date: str = None):
|
||||
"""
|
||||
初始化策略
|
||||
|
||||
Args:
|
||||
target_date: 目标日期，格式 YYYY-MM-DD，默认为今天
|
||||
"""
|
||||
self.target_date = target_date or datetime.now().strftime('%Y-%m-%d')
|
||||
self.date_helper = DateHelper()
|
||||
|
||||
# 计算应该使用的数据日期
|
||||
self.annual_date = self.date_helper.get_annual_date(self.target_date)
|
||||
self.quarter_date = self.date_helper.get_quarter_date(self.target_date)
|
||||
self.lifecycle_year = self.date_helper.get_lifecycle_year(self.target_date)
|
||||
|
||||
self.lifecycle_calculator = CompanyLifecycleFactor()
|
||||
self.financial_analyzer = FinancialIndicatorAnalyzer()
|
||||
DB_URL = "mysql+pymysql://fac_pattern:Chlry$%.8_app@192.168.16.153:3307/my_quant_db"
|
||||
self.distance_calculator = AverageDistanceFactor(DB_URL)
|
||||
|
||||
# MySQL连接
|
||||
self.mysql_engine = create_engine(
|
||||
DB_URL,
|
||||
pool_size=5,
|
||||
max_overflow=10,
|
||||
pool_recycle=3600
|
||||
)
|
||||
|
||||
# 结果表名
|
||||
self.result_table = 'tech_fundamental_factor_scores'
|
||||
|
||||
# 科技概念板块列表
|
||||
self.tech_concepts = [
|
||||
"5G概念", "物联网", "云计算", "边缘计算", "信息安全", "国产软件",
|
||||
"大数据", "数据中心", "芯片", "MCU芯片", "汽车芯片", "存储芯片",
|
||||
"人工智能", "AIGC概念", "ChatGPT概念", "CPO概念", "华为鸿蒙",
|
||||
"华为海思", "华为算力", "量子科技", "区块链", "数字货币", "工业互联",
|
||||
"操作系统", "光刻机", "第三代半导体", "元宇宙概念", "云游戏", "信创",
|
||||
"东数西算", "PCB概念", "先进封装", "EDA概念", "Web3概念", "数据确权",
|
||||
"数据要素", "数字水印", "工业软件", "6G概念", "时空大数据", "算力租赁",
|
||||
"光通信", "英伟达概念", "星闪概念", "液冷服务器", "多模态AI", "Sora概念",
|
||||
"AI手机PC", "铜缆高速连接", "车联网", "财税数字化", "智谱AI", "AI智能体",
|
||||
"DeepSeek概念", "AI医疗概念"
|
||||
]
|
||||
|
||||
def get_tech_stocks(self) -> pd.DataFrame:
|
||||
"""
|
||||
获取科技概念板块的股票列表
|
||||
|
||||
Returns:
|
||||
pd.DataFrame: 包含股票代码和名称的DataFrame
|
||||
"""
|
||||
try:
|
||||
# 构建查询条件
|
||||
concepts_str = "', '".join(self.tech_concepts)
|
||||
query = text(f"""
|
||||
SELECT DISTINCT gp_code as stock_code, gp_name as stock_name, bk_name as concept_name
|
||||
FROM gp_gnbk
|
||||
WHERE bk_name IN ('{concepts_str}')
|
||||
ORDER BY gp_code
|
||||
""")
|
||||
|
||||
with self.mysql_engine.connect() as conn:
|
||||
df = pd.read_sql(query, conn)
|
||||
|
||||
return df
|
||||
|
||||
except Exception as e:
|
||||
logger.error(f"获取科技概念股票失败: {str(e)}")
|
||||
return pd.DataFrame()
|
||||
|
||||
def filter_by_lifecycle(self, stock_codes: List[str]) -> Dict[str, List[str]]:
|
||||
"""
|
||||
根据企业生命周期筛选股票
|
||||
|
||||
Args:
|
||||
stock_codes: 股票代码列表
|
||||
|
||||
Returns:
|
||||
Dict: 包含成长期和成熟期股票的字典
|
||||
"""
|
||||
try:
|
||||
# 批量计算生命周期
|
||||
lifecycle_df = self.lifecycle_calculator.batch_calculate_lifecycle_factors(stock_codes, self.lifecycle_year)
|
||||
|
||||
# 筛选目标阶段的股票
|
||||
# 引入期(1)和成长期(2)合并为成长期，成熟期(3)保持不变
|
||||
growth_stage_stocks = lifecycle_df[
|
||||
lifecycle_df['stage_id'].isin([1, 2])
|
||||
]['stock_code'].tolist()
|
||||
|
||||
mature_stage_stocks = lifecycle_df[
|
||||
lifecycle_df['stage_id'] == 3
|
||||
]['stock_code'].tolist()
|
||||
|
||||
result = {
|
||||
'growth': growth_stage_stocks,
|
||||
'mature': mature_stage_stocks
|
||||
}
|
||||
|
||||
return result
|
||||
|
||||
except Exception as e:
|
||||
logger.error(f"生命周期筛选失败: {str(e)}")
|
||||
return {'growth': [], 'mature': []}
|
||||
|
||||
def calculate_distance_factors(self, growth_stocks: List[str], mature_stocks: List[str], days=20) -> Tuple[pd.DataFrame, pd.DataFrame]:
|
||||
"""
|
||||
分别计算成长期和成熟期股票的平均距离因子
|
||||
使用目标日期前N天的数据计算
|
||||
|
||||
Args:
|
||||
growth_stocks: 成长期股票列表
|
||||
mature_stocks: 成熟期股票列表
|
||||
days: 计算距离因子使用的天数（默认20天）
|
||||
|
||||
Returns:
|
||||
Tuple: (成长期距离因子DataFrame, 成熟期距离因子DataFrame)
|
||||
"""
|
||||
try:
|
||||
# 将目标日期转换为datetime对象
|
||||
target_datetime = datetime.strptime(self.target_date, '%Y-%m-%d')
|
||||
|
||||
growth_distance_df = pd.DataFrame()
|
||||
mature_distance_df = pd.DataFrame()
|
||||
|
||||
# 计算成长期股票距离因子
|
||||
if growth_stocks:
|
||||
growth_data = self.distance_calculator.get_stock_data(growth_stocks, days=days, end_date=target_datetime)
|
||||
if not growth_data.empty:
|
||||
growth_indicators = self.distance_calculator.calculate_technical_indicators(growth_data, days=days)
|
||||
growth_distance_df = self.distance_calculator.calculate_distance_factor(growth_indicators)
|
||||
|
||||
# 计算成熟期股票距离因子
|
||||
if mature_stocks:
|
||||
mature_data = self.distance_calculator.get_stock_data(mature_stocks, days=days, end_date=target_datetime)
|
||||
if not mature_data.empty:
|
||||
mature_indicators = self.distance_calculator.calculate_technical_indicators(mature_data, days=days)
|
||||
mature_distance_df = self.distance_calculator.calculate_distance_factor(mature_indicators)
|
||||
|
||||
return growth_distance_df, mature_distance_df
|
||||
|
||||
except Exception as e:
|
||||
logger.error(f"计算距离因子失败: {str(e)}")
|
||||
import traceback
|
||||
traceback.print_exc()
|
||||
return pd.DataFrame(), pd.DataFrame()
|
||||
|
||||
def calculate_common_factors(self, stock_codes: List[str]) -> pd.DataFrame:
|
||||
"""
|
||||
计算通用因子
|
||||
|
||||
Args:
|
||||
stock_codes: 股票代码列表
|
||||
|
||||
Returns:
|
||||
pd.DataFrame: 包含通用因子的DataFrame
|
||||
"""
|
||||
try:
|
||||
results = []
|
||||
|
||||
for stock_code in stock_codes:
|
||||
try:
|
||||
|
||||
factor_data = {'stock_code': stock_code}
|
||||
|
||||
# 1. 毛利率（使用季度数据）
|
||||
gross_margin = self.financial_analyzer.analyze_gross_profit_margin(stock_code, self.quarter_date)
|
||||
factor_data['gross_profit_margin'] = gross_margin
|
||||
|
||||
# 2. 成长能力指标
|
||||
growth_capability = self.financial_analyzer.analyze_growth_capability(stock_code)
|
||||
if growth_capability is not None:
|
||||
# 成长能力越高越好，使用sigmoid函数映射到0-1
|
||||
growth_score = 1 / (1 + math.exp(-growth_capability))
|
||||
else:
|
||||
growth_score = 0.5 # 默认中性评分
|
||||
factor_data['growth_score'] = growth_score
|
||||
|
||||
# 3. 前五大供应商占比（使用年报数据）
|
||||
supplier_conc = self.financial_analyzer.analyze_supplier_concentration(stock_code, self.annual_date)
|
||||
factor_data['supplier_concentration'] = supplier_conc
|
||||
|
||||
# 4. 前五大客户占比（使用年报数据）
|
||||
customer_conc = self.financial_analyzer.analyze_customer_concentration(stock_code, self.annual_date)
|
||||
factor_data['customer_concentration'] = customer_conc
|
||||
|
||||
results.append(factor_data)
|
||||
|
||||
except Exception as e:
|
||||
logger.warning(f"计算股票 {stock_code} 通用因子失败: {str(e)}")
|
||||
continue
|
||||
|
||||
df = pd.DataFrame(results)
|
||||
return df
|
||||
|
||||
except Exception as e:
|
||||
logger.error(f"计算通用因子失败: {str(e)}")
|
||||
return pd.DataFrame()
|
||||
|
||||
def calculate_growth_specific_factors(self, stock_codes: List[str]) -> pd.DataFrame:
|
||||
"""
|
||||
计算成长期特色因子
|
||||
|
||||
Args:
|
||||
stock_codes: 成长期股票代码列表
|
||||
|
||||
Returns:
|
||||
pd.DataFrame: 包含成长期特色因子的DataFrame
|
||||
"""
|
||||
try:
|
||||
results = []
|
||||
|
||||
for stock_code in stock_codes:
|
||||
try:
|
||||
|
||||
factor_data = {'stock_code': stock_code}
|
||||
|
||||
# 1. 管理费用率（使用季度数据）
|
||||
admin_ratio = self.financial_analyzer.analyze_admin_expense_ratio(stock_code, self.quarter_date)
|
||||
factor_data['admin_expense_ratio'] = admin_ratio
|
||||
|
||||
# 2. 研发费用折旧摊销占比（使用年度数据）
|
||||
financial_data = self.financial_analyzer.get_financial_data(stock_code, self.annual_date)
|
||||
if financial_data:
|
||||
intangible_amortize = financial_data.get('cash_flow_statement', {}).get('IA_AMORTIZE', 0)
|
||||
rd_expense = financial_data.get('profit_statement', {}).get('RESEARCH_EXPENSE', 0)
|
||||
|
||||
if rd_expense and rd_expense != 0:
|
||||
rd_amortize_ratio = intangible_amortize / rd_expense if intangible_amortize else 0
|
||||
else:
|
||||
rd_amortize_ratio = None # 使用None而不是0，避免这些股票获得最高分
|
||||
|
||||
factor_data['rd_amortize_ratio'] = rd_amortize_ratio
|
||||
else:
|
||||
factor_data['rd_amortize_ratio'] = None
|
||||
|
||||
# 3. 资产负债率（使用季度数据）
|
||||
asset_liability_ratio = self.financial_analyzer.analyze_asset_liability_ratio(stock_code, self.quarter_date)
|
||||
factor_data['asset_liability_ratio'] = asset_liability_ratio
|
||||
|
||||
results.append(factor_data)
|
||||
|
||||
except Exception as e:
|
||||
logger.warning(f"计算股票 {stock_code} 成长期特色因子失败: {str(e)}")
|
||||
continue
|
||||
|
||||
df = pd.DataFrame(results)
|
||||
return df
|
||||
|
||||
except Exception as e:
|
||||
logger.error(f"计算成长期特色因子失败: {str(e)}")
|
||||
return pd.DataFrame()
|
||||
|
||||
def calculate_mature_specific_factors(self, stock_codes: List[str]) -> pd.DataFrame:
|
||||
"""
|
||||
计算成熟期特色因子
|
||||
|
||||
Args:
|
||||
stock_codes: 成熟期股票代码列表
|
||||
|
||||
Returns:
|
||||
pd.DataFrame: 包含成熟期特色因子的DataFrame
|
||||
"""
|
||||
try:
|
||||
# 在循环外获取全A股PB和ROE数据，避免重复查询
|
||||
all_pb_data = self.financial_analyzer.get_all_stocks_pb_data()
|
||||
all_roe_data = self.financial_analyzer.get_all_stocks_roe_data(self.quarter_date)
|
||||
|
||||
results = []
|
||||
|
||||
for stock_code in stock_codes:
|
||||
try:
|
||||
factor_data = {'stock_code': stock_code}
|
||||
|
||||
# 1. 应收账款周转率（使用季度数据）
|
||||
formatted_stock_code = self.financial_analyzer.code_formatter.to_dot_format(stock_code)
|
||||
financial_data = self.financial_analyzer.get_financial_data(formatted_stock_code, self.quarter_date)
|
||||
if financial_data:
|
||||
revenue = financial_data.get('profit_statement', {}).get('OPERATE_INCOME', 0)
|
||||
accounts_rece = financial_data.get('balance_sheet', {}).get('ACCOUNTS_RECE', 0)
|
||||
|
||||
if accounts_rece and accounts_rece != 0:
|
||||
turnover_ratio = revenue / accounts_rece if revenue else 0
|
||||
else:
|
||||
turnover_ratio = None # 使用None而不是0
|
||||
|
||||
factor_data['accounts_receivable_turnover'] = turnover_ratio
|
||||
else:
|
||||
factor_data['accounts_receivable_turnover'] = None
|
||||
|
||||
# 2. 研发强度（使用季度数据）
|
||||
rd_intensity = self.financial_analyzer.analyze_rd_expense_ratio(stock_code, self.quarter_date)
|
||||
factor_data['rd_intensity'] = rd_intensity
|
||||
|
||||
# 3. PB-ROE排名因子：使用预获取的全A股数据
|
||||
if all_pb_data and all_roe_data:
|
||||
pb_roe_rank_factor = self.financial_analyzer.calculate_pb_roe_rank_factor(
|
||||
stock_code, all_pb_data, all_roe_data
|
||||
)
|
||||
factor_data['pb_roe_rank_factor'] = pb_roe_rank_factor
|
||||
else:
|
||||
factor_data['pb_roe_rank_factor'] = None
|
||||
|
||||
results.append(factor_data)
|
||||
|
||||
except Exception as e:
|
||||
logger.warning(f"计算股票 {stock_code} 成熟期特色因子失败: {str(e)}")
|
||||
continue
|
||||
|
||||
df = pd.DataFrame(results)
|
||||
return df
|
||||
|
||||
except Exception as e:
|
||||
logger.error(f"计算成熟期特色因子失败: {str(e)}")
|
||||
return pd.DataFrame()
|
||||
|
||||
def run_strategy(self) -> Dict[str, pd.DataFrame]:
|
||||
"""
|
||||
运行完整的选股策略
|
||||
|
||||
Returns:
|
||||
Dict: 包含成长期和成熟期股票分析结果的字典
|
||||
"""
|
||||
try:
|
||||
# 1. 获取科技概念股票
|
||||
tech_stocks_df = self.get_tech_stocks()
|
||||
if tech_stocks_df.empty:
|
||||
logger.error("未获取到科技概念股票")
|
||||
return {}
|
||||
|
||||
stock_codes = tech_stocks_df['stock_code'].unique().tolist()
|
||||
|
||||
# 2. 按企业生命周期筛选
|
||||
lifecycle_result = self.filter_by_lifecycle(stock_codes)
|
||||
growth_stocks = lifecycle_result['growth']
|
||||
mature_stocks = lifecycle_result['mature']
|
||||
|
||||
if not growth_stocks and not mature_stocks:
|
||||
logger.warning("未找到符合条件的成长期或成熟期股票")
|
||||
return {}
|
||||
|
||||
# 3. 计算平均距离因子
|
||||
growth_distance_df, mature_distance_df = self.calculate_distance_factors(growth_stocks, mature_stocks)
|
||||
|
||||
# 4. 计算通用因子
|
||||
all_qualified_stocks = growth_stocks + mature_stocks
|
||||
common_factors_df = self.calculate_common_factors(all_qualified_stocks)
|
||||
|
||||
# 5. 计算特色因子
|
||||
growth_specific_df = self.calculate_growth_specific_factors(growth_stocks) if growth_stocks else pd.DataFrame()
|
||||
mature_specific_df = self.calculate_mature_specific_factors(mature_stocks) if mature_stocks else pd.DataFrame()
|
||||
|
||||
# 6. 合并结果并计算分数
|
||||
result = {}
|
||||
|
||||
# 处理成长期股票
|
||||
if not growth_specific_df.empty:
|
||||
# 成长期结果合并
|
||||
growth_result = growth_specific_df.copy()
|
||||
|
||||
# 合并距离因子
|
||||
if not growth_distance_df.empty:
|
||||
growth_result = growth_result.merge(
|
||||
growth_distance_df[['symbol', 'avg_distance_factor']],
|
||||
left_on='stock_code', right_on='symbol', how='left'
|
||||
).drop('symbol', axis=1)
|
||||
|
||||
# 合并通用因子
|
||||
if not common_factors_df.empty:
|
||||
growth_result = growth_result.merge(
|
||||
common_factors_df, on='stock_code', how='left'
|
||||
)
|
||||
|
||||
# 计算因子分数
|
||||
growth_result = self.calculate_factor_scores(growth_result, 'growth')
|
||||
|
||||
# 计算总分并排序
|
||||
growth_result = self.calculate_total_score(growth_result, 'growth')
|
||||
|
||||
result['growth'] = growth_result
|
||||
|
||||
# 处理成熟期股票
|
||||
if not mature_specific_df.empty:
|
||||
# 成熟期结果合并
|
||||
mature_result = mature_specific_df.copy()
|
||||
|
||||
# 合并距离因子
|
||||
if not mature_distance_df.empty:
|
||||
mature_result = mature_result.merge(
|
||||
mature_distance_df[['symbol', 'avg_distance_factor']],
|
||||
left_on='stock_code', right_on='symbol', how='left'
|
||||
).drop('symbol', axis=1)
|
||||
|
||||
# 合并通用因子
|
||||
if not common_factors_df.empty:
|
||||
mature_result = mature_result.merge(
|
||||
common_factors_df, on='stock_code', how='left'
|
||||
)
|
||||
|
||||
# 计算因子分数
|
||||
mature_result = self.calculate_factor_scores(mature_result, 'mature')
|
||||
|
||||
# 计算总分并排序
|
||||
mature_result = self.calculate_total_score(mature_result, 'mature')
|
||||
|
||||
result['mature'] = mature_result
|
||||
|
||||
# 7. 保存结果到数据库
|
||||
if result:
|
||||
self.save_results_to_db(result)
|
||||
|
||||
return result
|
||||
|
||||
except Exception as e:
|
||||
logger.error(f"策略运行失败: {str(e)}")
|
||||
return {}
|
||||
|
||||
def calculate_factor_scores(self, df: pd.DataFrame, stage: str) -> pd.DataFrame:
|
||||
"""
|
||||
计算单因子打分（0-100分位数）
|
||||
|
||||
Args:
|
||||
df: 包含因子数据的DataFrame
|
||||
stage: 阶段类型 ('growth' 或 'mature')
|
||||
|
||||
Returns:
|
||||
pd.DataFrame: 包含因子分数的DataFrame
|
||||
"""
|
||||
try:
|
||||
if df.empty:
|
||||
return df
|
||||
|
||||
df_scored = df.copy()
|
||||
|
||||
# 定义因子方向（正向为True，负向为False）
|
||||
factor_directions = {
|
||||
# 通用因子
|
||||
'gross_profit_margin': True, # 毛利率_环比增量 - 正向
|
||||
'growth_score': True, # 成长能力 - 正向
|
||||
'supplier_concentration': False, # 前5大供应商金额占比合计 - 负向
|
||||
'customer_concentration': False, # 前5大客户收入金额占比合计 - 负向
|
||||
'avg_distance_factor': False, # 平均距离因子 - 负向
|
||||
|
||||
# 成长期特色因子
|
||||
'admin_expense_ratio': False, # 管理费用/营业总收入_环比增量 - 负向
|
||||
'rd_amortize_ratio': False, # 研发费用折旧摊销占比_环比增量 - 负向
|
||||
'asset_liability_ratio': True, # 资产负债率 - 正向
|
||||
|
||||
# 成熟期特色因子
|
||||
'accounts_receivable_turnover': True, # 应收账款周转率 - 正向
|
||||
'rd_intensity': True, # 研发费用直接投入占比_环比增量 - 正向
|
||||
'pb_roe_rank_factor': False # PB-ROE排名因子 - 负向（越小越好）
|
||||
}
|
||||
|
||||
# 为每个因子计算分位数分数
|
||||
for column in df.columns:
|
||||
if column == 'stock_code':
|
||||
continue
|
||||
|
||||
# 只对有效值进行排名计算
|
||||
values = df_scored[column].dropna()
|
||||
if len(values) <= 1:
|
||||
# 如果只有一个值或没有值，所有股票都得50分或0分
|
||||
if len(values) == 1:
|
||||
df_scored[f'{column}_score'] = df_scored[column].apply(lambda x: 50 if pd.notna(x) else 0)
|
||||
else:
|
||||
df_scored[f'{column}_score'] = 0
|
||||
continue
|
||||
|
||||
# 根据因子方向确定排序方式
|
||||
is_positive = factor_directions.get(column, True)
|
||||
|
||||
# 计算排名分数
|
||||
if is_positive:
|
||||
# 正向因子：值越大分数越高
|
||||
ranked_values = values.rank(pct=True) * 100
|
||||
else:
|
||||
# 负向因子：值越小分数越高
|
||||
ranked_values = (1 - values.rank(pct=True)) * 100
|
||||
|
||||
# 初始化分数列
|
||||
df_scored[f'{column}_score'] = 0.0
|
||||
|
||||
# 将分数赋值给对应的行
|
||||
for idx in ranked_values.index:
|
||||
df_scored.loc[idx, f'{column}_score'] = ranked_values[idx]
|
||||
|
||||
return df_scored
|
||||
|
||||
except Exception as e:
|
||||
logger.error(f"计算因子分数失败: {str(e)}")
|
||||
import traceback
|
||||
traceback.print_exc()
|
||||
return df
|
||||
|
||||
def calculate_total_score(self, df: pd.DataFrame, stage: str) -> pd.DataFrame:
|
||||
"""
|
||||
计算总分
|
||||
使用公式：总分 = 1/8 * Mean(Si) + Mean(Si)/Std(Si)
|
||||
|
||||
Args:
|
||||
df: 包含因子分数的DataFrame
|
||||
stage: 阶段类型 ('growth' 或 'mature')
|
||||
|
||||
Returns:
|
||||
pd.DataFrame: 包含总分的DataFrame
|
||||
"""
|
||||
try:
|
||||
if df.empty:
|
||||
return df
|
||||
|
||||
df_result = df.copy()
|
||||
|
||||
# 定义因子权重（注意：这里是factor_score而不是factor）
|
||||
if stage == 'growth':
|
||||
factor_weights = {
|
||||
# 通用因子
|
||||
'gross_profit_margin_score': 1/8,
|
||||
'growth_score_score': 1/8, # 注意这里是growth_score_score
|
||||
'supplier_concentration_score': 1/8,
|
||||
'customer_concentration_score': 1/8,
|
||||
'avg_distance_factor_score': 1/8,
|
||||
|
||||
# 成长期特色因子
|
||||
'admin_expense_ratio_score': 1/8,
|
||||
'rd_amortize_ratio_score': 1/8,
|
||||
'asset_liability_ratio_score': 1/8
|
||||
}
|
||||
else: # mature
|
||||
factor_weights = {
|
||||
# 通用因子
|
||||
'gross_profit_margin_score': 1/8,
|
||||
'growth_score_score': 1/8, # 注意这里是growth_score_score
|
||||
'supplier_concentration_score': 1/8,
|
||||
'customer_concentration_score': 1/8,
|
||||
'avg_distance_factor_score': 1/8,
|
||||
|
||||
# 成熟期特色因子
|
||||
'accounts_receivable_turnover_score': 1/8,
|
||||
'rd_intensity_score': 1/8,
|
||||
'pb_roe_rank_factor_score': 1/8
|
||||
}
|
||||
|
||||
# 计算每只股票的总分
|
||||
total_scores = []
|
||||
|
||||
for index, row in df_result.iterrows():
|
||||
# 获取该股票的所有因子分数
|
||||
factor_scores = []
|
||||
valid_weights = []
|
||||
|
||||
for factor, weight in factor_weights.items():
|
||||
if factor in row and pd.notna(row[factor]) and row[factor] > 0:
|
||||
factor_scores.append(row[factor])
|
||||
valid_weights.append(weight)
|
||||
|
||||
if len(factor_scores) == 0:
|
||||
total_scores.append(0)
|
||||
continue
|
||||
|
||||
factor_scores = np.array(factor_scores)
|
||||
valid_weights = np.array(valid_weights)
|
||||
|
||||
# 重新标准化权重
|
||||
valid_weights = valid_weights / valid_weights.sum()
|
||||
|
||||
# 计算加权平均分数
|
||||
mean_score = np.average(factor_scores, weights=valid_weights)
|
||||
|
||||
# 计算调整项 Mean(Si)/Std(Si)
|
||||
if len(factor_scores) > 1 and np.std(factor_scores) > 0:
|
||||
adjustment = np.mean(factor_scores) / np.std(factor_scores)
|
||||
else:
|
||||
adjustment = 0
|
||||
|
||||
# 计算总分：1/8 * Mean(Si) + Mean(Si)/Std(Si)
|
||||
total_score = (1/8) * mean_score + adjustment
|
||||
total_scores.append(total_score)
|
||||
|
||||
df_result['total_score'] = total_scores
|
||||
|
||||
# 按总分降序排列
|
||||
df_result = df_result.sort_values('total_score', ascending=False).reset_index(drop=True)
|
||||
df_result['rank'] = range(1, len(df_result) + 1)
|
||||
|
||||
return df_result
|
||||
|
||||
except Exception as e:
|
||||
logger.error(f"计算总分失败: {str(e)}")
|
||||
import traceback
|
||||
traceback.print_exc()
|
||||
return df
|
||||
|
||||
def save_results_to_db(self, results: Dict[str, pd.DataFrame]) -> bool:
|
||||
"""
|
||||
将策略结果保存到数据库
|
||||
|
||||
Args:
|
||||
results: 策略结果字典，包含 'growth' 和 'mature' 两个DataFrame
|
||||
|
||||
Returns:
|
||||
bool: 是否保存成功
|
||||
"""
|
||||
try:
|
||||
if not results:
|
||||
return False
|
||||
|
||||
# 准备入库数据
|
||||
records = []
|
||||
for stage, df in results.items():
|
||||
if df.empty or 'stock_code' not in df.columns or 'total_score' not in df.columns:
|
||||
continue
|
||||
|
||||
for _, row in df.iterrows():
|
||||
stock_code = row['stock_code']
|
||||
total_score = row.get('total_score', 0)
|
||||
|
||||
# 跳过总分为空或0的记录
|
||||
if pd.isna(total_score) or total_score == 0:
|
||||
continue
|
||||
|
||||
records.append({
|
||||
'trade_date': self.target_date,
|
||||
'stock_code': stock_code,
|
||||
'stage': stage, # 'growth' 或 'mature'
|
||||
'total_score': float(total_score),
|
||||
'created_at': datetime.now(),
|
||||
'updated_at': datetime.now()
|
||||
})
|
||||
|
||||
if not records:
|
||||
return False
|
||||
|
||||
# 删除该日期的旧数据
|
||||
with self.mysql_engine.begin() as conn:
|
||||
delete_sql = text(f"""
|
||||
DELETE FROM {self.result_table}
|
||||
WHERE trade_date = :trade_date
|
||||
""")
|
||||
conn.execute(delete_sql, {"trade_date": self.target_date})
|
||||
|
||||
# 批量插入新数据
|
||||
if records:
|
||||
df_to_save = pd.DataFrame(records)
|
||||
df_to_save.to_sql(
|
||||
self.result_table,
|
||||
self.mysql_engine,
|
||||
if_exists='append',
|
||||
index=False
|
||||
)
|
||||
|
||||
return True
|
||||
|
||||
except Exception as e:
|
||||
logger.error(f"保存策略结果到数据库失败: {str(e)}")
|
||||
import traceback
|
||||
traceback.print_exc()
|
||||
return False
|
||||
|
||||
def close_connections(self):
|
||||
"""关闭所有数据库连接"""
|
||||
try:
|
||||
if hasattr(self, 'lifecycle_calculator'):
|
||||
del self.lifecycle_calculator
|
||||
if hasattr(self, 'financial_analyzer'):
|
||||
self.financial_analyzer.close_connection()
|
||||
if hasattr(self, 'distance_calculator'):
|
||||
del self.distance_calculator
|
||||
if hasattr(self, 'mysql_engine'):
|
||||
self.mysql_engine.dispose()
|
||||
except Exception as e:
|
||||
logger.error(f"关闭连接失败: {str(e)}")
|
||||
|
||||
|
||||
def main():
|
||||
"""主函数 - 科技主题基本面因子选股策略"""
|
||||
strategy = None
|
||||
try:
|
||||
# 创建策略实例（可以指定日期，默认为今天）
|
||||
import sys
|
||||
if len(sys.argv) > 1:
|
||||
target_date = sys.argv[1] # 从命令行参数获取日期
|
||||
else:
|
||||
target_date = "2025-09-01" # 默认日期，可以修改
|
||||
|
||||
strategy = TechFundamentalFactorStrategy(target_date=target_date)
|
||||
|
||||
# 运行策略（结果会自动保存到数据库）
|
||||
results = strategy.run_strategy()
|
||||
|
||||
except Exception as e:
|
||||
logger.error(f"程序执行失败: {str(e)}")
|
||||
import traceback
|
||||
traceback.print_exc()
|
||||
finally:
|
||||
if strategy:
|
||||
strategy.close_connections()
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
main()
|
||||
Loading…
Reference in New Issue