diff --git a/manage-instances.sh b/manage-instances.sh index 59bb893..3c0dbed 100644 --- a/manage-instances.sh +++ b/manage-instances.sh @@ -11,7 +11,8 @@ show_help() { echo " start [实例ID] 启动指定实例或所有实例" echo " stop [实例ID] 停止指定实例或所有实例" echo " restart [实例ID] 重启指定实例或所有实例" - echo " logs [实例ID] 实时查看指定实例的日志 (Ctrl+C 退出)" + echo " logs [实例ID] 查看最新10条日志并实时跟踪新日志" + echo " logs-follow [实例ID] 实时查看指定实例的日志 (Ctrl+C 退出)" echo " status 显示实例状态概览" echo " remove [实例ID] 删除指定实例或所有实例" echo " rebuild [数量] 重新构建镜像并部署指定数量的实例" @@ -21,7 +22,8 @@ show_help() { echo " $0 list 列出所有实例" echo " $0 start 2 启动实例2" echo " $0 stop all 停止所有实例" - echo " $0 logs 1 查看实例1的日志" + echo " $0 logs 1 查看实例1最新10条日志并实时跟踪" + echo " $0 logs-follow 1 实时跟踪实例1的日志" echo " $0 rebuild 2 重新构建并部署2个实例" echo " $0 update 热更新所有实例的代码" exit 1 @@ -68,6 +70,13 @@ restart_instance() { # 函数：查看实例日志 view_logs() { + echo "显示实例 $1 的最新10条日志，然后实时跟踪新日志 (按 Ctrl+C 退出):" + echo "----------------------------------------" + docker logs --tail 10 -f stock-app-$1 +} + +# 函数：实时查看实例日志 +view_logs_follow() { echo "正在实时显示实例 $1 的日志 (按 Ctrl+C 退出):" echo "----------------------------------------" docker logs -f stock-app-$1 @@ -187,6 +196,13 @@ case "$1" in fi view_logs $2 ;; + logs-follow) + if [ "$#" -lt 2 ]; then + echo "错误: 缺少实例ID参数" + show_help + fi + view_logs_follow $2 + ;; status) show_status ;; diff --git a/src/QMT/strategy.py b/src/QMT/strategy.py index 389f4bb..b3e28d1 100644 --- a/src/QMT/strategy.py +++ b/src/QMT/strategy.py @@ -245,7 +245,7 @@ def create_buy_strategy_callback(xt_trader, acc, buy_amount, logger): # 检查是否有在途订单（Redis） if is_stock_pending_order(stock_code): - logger.info(f"{stock_code} 有在途订单，跳过买入") + # logger.info(f"{stock_code} 有在途订单，跳过买入") continue # 集合竞价时段：只观察，不下单 @@ -339,7 +339,7 @@ def create_sell_strategy_callback(xt_trader, acc, logger): # 检查是否有在途订单 if is_stock_pending_order(stock_code): - logger.info(f"{stock_code} 有在途订单，跳过卖出") + # logger.info(f"{stock_code} 有在途订单，跳过卖出") continue # 集合竞价时段：只观察，不下单 diff --git a/src/app.py b/src/app.py index ae1288b..155ae61 100644 --- a/src/app.py +++ b/src/app.py @@ -3404,6 +3404,74 @@ def analyze_valuation_indicator(): "message": f"估值指标分析失败: {str(e)}" }), 500 + +@app.route('/api/overlap/analysis', methods=['GET']) +def analyze_stock_overlap(): + """分析股票重叠度和滞涨情况 + GET参数: + - stock_code: 股票代码 (例如: 300661.SZ) + - days: 分析天数，默认3个交易日 + + 返回格式: + { + "status": "success", + "data": { + "target_stock": "300661.SZ", + "target_stock_name": "圣邦股份", + "analysis_date": "2025-09-28 15:30:42", + "analysis_period_days": 3, + "industries": [...], + "concepts": [...], + "overlap_threshold": 5, + "top5_stocks": [...], + "lag_analysis": {...} + } + } + """ + try: + # 获取参数 + stock_code = request.args.get('stock_code') + days = int(request.args.get('days', 3)) + + if not stock_code: + return jsonify({ + "status": "error", + "message": "缺少必需参数: stock_code" + }), 400 + + # 导入重叠度分析器 + from src.quantitative_analysis.overlap_analyzer import OverlapAnalyzer + + # 创建分析器实例 + analyzer = OverlapAnalyzer() + + try: + # 执行分析 + result = analyzer.analyze_stock_overlap(stock_code, days) + + if 'error' in result: + return jsonify({ + "status": "error", + "message": result['error'] + }), 500 + + return jsonify({ + "status": "success", + "data": result + }) + + finally: + # 关闭数据库连接 + analyzer.close_connection() + + except Exception as e: + logger.error(f"重叠度分析失败: {str(e)}") + return jsonify({ + "status": "error", + "message": f"重叠度分析失败: {str(e)}" + }), 500 + + if __name__ == '__main__': # 启动Web服务器 diff --git a/src/fundamentals_llm/fonts/simhei.ttf b/src/fundamentals_llm/fonts/simhei.ttf new file mode 100644 index 0000000..3326815 Binary files /dev/null and b/src/fundamentals_llm/fonts/simhei.ttf differ diff --git a/src/quantitative_analysis/average_distance_factor.py b/src/quantitative_analysis/average_distance_factor.py index 4b563ab..a6be34a 100644 --- a/src/quantitative_analysis/average_distance_factor.py +++ b/src/quantitative_analysis/average_distance_factor.py @@ -289,10 +289,10 @@ def main(): analyzer = AverageDistanceFactor(db_url) # 示例1: 分析特定行业 - # result = analyzer.analyze_industry(industry_name="旅游") + result = analyzer.analyze_industry(industry_name="军工电子") # 示例2: 分析特定概念 - result = analyzer.analyze_industry(concept_name="人形机器人") + # result = analyzer.analyze_industry(concept_name="固态电池") # 示例3: 查看可用的行业列表 # industries = analyzer.get_available_industries() diff --git a/src/quantitative_analysis/overlap_analyzer.py b/src/quantitative_analysis/overlap_analyzer.py new file mode 100644 index 0000000..77249fa --- /dev/null +++ b/src/quantitative_analysis/overlap_analyzer.py @@ -0,0 +1,668 @@ +#!/usr/bin/env python3 +# -*- coding: utf-8 -*- + +""" +行业重叠度和概念重叠度分析工具 +用于分析某股票与相似股票的重叠度，识别滞涨投资机会 +""" + +import sys +import os +import logging +from typing import Dict, List, Optional, Tuple +from pathlib import Path +from sqlalchemy import create_engine, text +import pandas as pd +from datetime import datetime, timedelta + +# 添加项目根路径到Python路径 +project_root = Path(__file__).parent.parent.parent +sys.path.append(str(project_root)) + +# 导入配置 +try: + from src.valuation_analysis.config import DB_URL +except ImportError: + # 如果上面的导入失败，尝试直接导入 + import importlib.util + config_path = os.path.join(project_root, 'src', 'valuation_analysis', 'config.py') + spec = importlib.util.spec_from_file_location("config", config_path) + config_module = importlib.util.module_from_spec(spec) + spec.loader.exec_module(config_module) + DB_URL = config_module.DB_URL + +# 导入股票代码格式转换工具 +try: + from tools.stock_code_formatter import StockCodeFormatter +except ImportError: + # 如果上面的导入失败，尝试直接导入 + import importlib.util + formatter_path = os.path.join(project_root, 'tools', 'stock_code_formatter.py') + spec = importlib.util.spec_from_file_location("stock_code_formatter", formatter_path) + formatter_module = importlib.util.module_from_spec(spec) + spec.loader.exec_module(formatter_module) + StockCodeFormatter = formatter_module.StockCodeFormatter + +# 设置日志 +logging.basicConfig( + level=logging.INFO, + format='%(asctime)s - %(name)s - %(levelname)s - %(message)s' +) +logger = logging.getLogger(__name__) + + +class OverlapAnalyzer: + """行业重叠度和概念重叠度分析器""" + + def __init__(self): + """初始化""" + # MySQL连接 + self.mysql_engine = None + + # 股票代码格式转换器 + self.code_formatter = StockCodeFormatter() + + # 重叠度评分权重 + self.INDUSTRY_WEIGHT = 3 # 行业重叠权重 + self.CONCEPT_WEIGHT = 1 # 概念重叠权重 + self.MIN_OVERLAP_SCORE = 5 # 最小重叠分数阈值 + + self.connect_mysql() + + def connect_mysql(self): + """连接MySQL数据库""" + try: + self.mysql_engine = create_engine( + DB_URL, + pool_size=5, + max_overflow=10, + pool_recycle=3600 + ) + + # 测试连接 + with self.mysql_engine.connect() as conn: + conn.execute(text("SELECT 1")) + + logger.info("MySQL数据库连接成功") + + except Exception as e: + logger.error(f"MySQL数据库连接失败: {str(e)}") + raise + + def normalize_stock_code(self, stock_code: str) -> str: + """ + 标准化股票代码格式，转换为数据库中使用的格式 + + Args: + stock_code: 输入的股票代码，支持多种格式 + + Returns: + str: 标准化后的股票代码 + """ + return self.code_formatter.to_prefix_format(stock_code) + + def get_stock_industries(self, stock_code: str) -> List[Tuple[str, str]]: + """ + 获取股票所属的行业板块 + + Args: + stock_code: 股票代码 + + Returns: + List[Tuple[str, str]]: [(板块代码, 板块名称)] 列表 + """ + try: + normalized_code = self.normalize_stock_code(stock_code) + + query = text(""" + SELECT DISTINCT bk_code, bk_name + FROM gp_hybk + WHERE gp_code = :stock_code + AND bk_code IS NOT NULL + AND bk_name IS NOT NULL + """) + + with self.mysql_engine.connect() as conn: + result = conn.execute(query, {"stock_code": normalized_code}).fetchall() + + industries = [(str(row[0]), str(row[1])) for row in result] + logger.info(f"股票 {stock_code} 找到 {len(industries)} 个行业板块") + return industries + + except Exception as e: + logger.error(f"获取股票 {stock_code} 行业板块失败: {str(e)}") + return [] + + def get_stock_concepts(self, stock_code: str) -> List[Tuple[str, str]]: + """ + 获取股票所属的概念板块 + + Args: + stock_code: 股票代码 + + Returns: + List[Tuple[str, str]]: [(板块代码, 板块名称)] 列表 + """ + try: + normalized_code = self.normalize_stock_code(stock_code) + + query = text(""" + SELECT DISTINCT bk_code, bk_name + FROM gp_gnbk + WHERE gp_code = :stock_code + AND bk_code IS NOT NULL + AND bk_name IS NOT NULL + """) + + with self.mysql_engine.connect() as conn: + result = conn.execute(query, {"stock_code": normalized_code}).fetchall() + + concepts = [(str(row[0]), str(row[1])) for row in result] + logger.info(f"股票 {stock_code} 找到 {len(concepts)} 个概念板块") + return concepts + + except Exception as e: + logger.error(f"获取股票 {stock_code} 概念板块失败: {str(e)}") + return [] + + def get_similar_stocks_by_industry(self, industries: List[Tuple[str, str]]) -> Dict[str, List[str]]: + """ + 根据行业板块获取相似股票 + + Args: + industries: 行业板块列表 [(板块代码, 板块名称)] + + Returns: + Dict[str, List[str]]: {板块代码: [股票代码列表]} + """ + try: + similar_stocks = {} + + for bk_code, bk_name in industries: + query = text(""" + SELECT DISTINCT gp_code + FROM gp_hybk + WHERE bk_code = :bk_code + AND gp_code IS NOT NULL + """) + + with self.mysql_engine.connect() as conn: + result = conn.execute(query, {"bk_code": bk_code}).fetchall() + + stock_codes = [row[0] for row in result if row[0]] + similar_stocks[bk_code] = stock_codes + logger.info(f"行业板块 {bk_name}({bk_code}) 包含 {len(stock_codes)} 只股票") + + return similar_stocks + + except Exception as e: + logger.error(f"获取行业相似股票失败: {str(e)}") + return {} + + def get_similar_stocks_by_concept(self, concepts: List[Tuple[str, str]]) -> Dict[str, List[str]]: + """ + 根据概念板块获取相似股票 + + Args: + concepts: 概念板块列表 [(板块代码, 板块名称)] + + Returns: + Dict[str, List[str]]: {板块代码: [股票代码列表]} + """ + try: + similar_stocks = {} + + for bk_code, bk_name in concepts: + query = text(""" + SELECT DISTINCT gp_code + FROM gp_gnbk + WHERE bk_code = :bk_code + AND gp_code IS NOT NULL + """) + + with self.mysql_engine.connect() as conn: + result = conn.execute(query, {"bk_code": bk_code}).fetchall() + + stock_codes = [row[0] for row in result if row[0]] + similar_stocks[bk_code] = stock_codes + logger.info(f"概念板块 {bk_name}({bk_code}) 包含 {len(stock_codes)} 只股票") + + return similar_stocks + + except Exception as e: + logger.error(f"获取概念相似股票失败: {str(e)}") + return {} + + def calculate_overlap_scores(self, target_stock: str, industry_stocks: Dict[str, List[str]], concept_stocks: Dict[str, List[str]]) -> Dict[str, float]: + """ + 计算重叠度分数 + + Args: + target_stock: 目标股票代码 + industry_stocks: 行业相似股票字典 {板块代码: [股票代码列表]} + concept_stocks: 概念相似股票字典 {板块代码: [股票代码列表]} + + Returns: + Dict[str, float]: {股票代码: 重叠分数} + """ + try: + overlap_scores = {} + target_stock_normalized = self.normalize_stock_code(target_stock) + + # 计算行业重叠分数（每个行业+3分） + for bk_code, stock_list in industry_stocks.items(): + for stock_code in stock_list: + if stock_code == target_stock_normalized: + continue # 跳过目标股票本身 + + if stock_code not in overlap_scores: + overlap_scores[stock_code] = 0 + + overlap_scores[stock_code] += self.INDUSTRY_WEIGHT + + # 计算概念重叠分数（每个概念+1分） + for bk_code, stock_list in concept_stocks.items(): + for stock_code in stock_list: + if stock_code == target_stock_normalized: + continue # 跳过目标股票本身 + + if stock_code not in overlap_scores: + overlap_scores[stock_code] = 0 + + overlap_scores[stock_code] += self.CONCEPT_WEIGHT + + logger.info(f"计算出 {len(overlap_scores)} 只股票的重叠分数") + return overlap_scores + + except Exception as e: + logger.error(f"计算重叠分数失败: {str(e)}") + return {} + + def filter_high_overlap_stocks(self, overlap_scores: Dict[str, float]) -> List[Tuple[str, float]]: + """ + 筛选高重叠度股票 + + Args: + overlap_scores: 重叠分数字典 + + Returns: + List[Tuple[str, float]]: [(股票代码, 分数)] 按分数降序排列 + """ + try: + # 筛选分数大于阈值的股票 + high_overlap_stocks = [ + (stock_code, score) for stock_code, score in overlap_scores.items() + if score > self.MIN_OVERLAP_SCORE + ] + + # 按分数降序排列 + high_overlap_stocks.sort(key=lambda x: x[1], reverse=True) + + logger.info(f"筛选出 {len(high_overlap_stocks)} 只高重叠度股票（分数>{self.MIN_OVERLAP_SCORE}）") + return high_overlap_stocks + + except Exception as e: + logger.error(f"筛选高重叠度股票失败: {str(e)}") + return [] + + def get_recent_price_changes(self, stock_codes: List[str], days: int = 3) -> Dict[str, Dict]: + """ + 获取股票的近期涨跌幅（默认近3个交易日） + + Args: + stock_codes: 股票代码列表 + days: 统计天数，默认3个交易日 + + Returns: + Dict[str, Dict]: {股票代码: {price_change, change_pct, latest_price}} + """ + try: + price_changes = {} + + for stock_code in stock_codes: + # 获取最近N个交易日的价格数据 + query = text(""" + SELECT close, percent, timestamp + FROM gp_day_data + WHERE symbol = :stock_code + ORDER BY timestamp DESC + LIMIT :days + """) + + with self.mysql_engine.connect() as conn: + result = conn.execute(query, {"stock_code": stock_code, "days": days}).fetchall() + + if len(result) >= 2: # 至少需要2个交易日的数据 + # 最新价格（第一个） + latest_price = float(result[0][0]) if result[0][0] else None + latest_change_pct = float(result[0][1]) if result[0][1] else None + + # N天前价格（最后一个） + historical_price = float(result[-1][0]) if result[-1][0] else None + + if latest_price and historical_price: + price_change = latest_price - historical_price + change_pct = (price_change / historical_price) * 100 + + price_changes[stock_code] = { + 'latest_price': latest_price, + 'price_change': price_change, + 'change_pct': change_pct, + 'latest_change_pct': latest_change_pct, + 'trading_days': len(result) + } + + logger.info(f"获取到 {len(price_changes)} 只股票的近{days}个交易日涨跌幅数据") + return price_changes + + except Exception as e: + logger.error(f"获取近期涨跌幅失败: {str(e)}") + return {} + + def analyze_lag_performance(self, target_change: Dict, top5_stocks: List[Tuple[str, float]], similar_changes: Dict[str, Dict]) -> Dict: + """ + 分析目标股票的滞涨情况（基于近3个交易日） + + Args: + target_change: 目标股票涨跌幅数据 + top5_stocks: 重叠度最高的5只股票 [(股票代码, 重叠分数)] + similar_changes: 相似股票涨跌幅数据 + + Returns: + Dict: 滞涨分析结果 + """ + try: + if not target_change or 'change_pct' not in target_change: + logger.warning("目标股票涨跌幅数据不完整，无法进行滞涨分析") + return { + 'target_lag_vs_top5': None, + 'target_lag_vs_top1': None, + 'top5_avg_change': None, + 'top1_change': None + } + + target_change_pct = target_change['change_pct'] + + # 获取前5只股票的有效涨跌幅数据 + top5_changes = [] + for stock_code, _ in top5_stocks: + if stock_code in similar_changes and 'change_pct' in similar_changes[stock_code]: + change_pct = similar_changes[stock_code]['change_pct'] + if change_pct is not None: + top5_changes.append(change_pct) + + if len(top5_changes) < 1: + logger.warning("没有有效的前5只股票涨跌幅数据") + return { + 'target_lag_vs_top5': None, + 'target_lag_vs_top1': None, + 'top5_avg_change': None, + 'top1_change': None + } + + # 计算前5只股票的平均涨幅 + top5_avg_change = sum(top5_changes) / len(top5_changes) + + # 获取重叠度最高的1只股票涨幅 + top1_change = top5_changes[0] if top5_changes else None + + # 计算滞涨分数 + lag_vs_top5 = target_change_pct - top5_avg_change if top5_avg_change is not None else None + lag_vs_top1 = target_change_pct - top1_change if top1_change is not None else None + + # 滞涨等级判断 + def get_lag_level(lag_score): + if lag_score is None: + return "无数据" + elif lag_score < -8: + return "严重滞涨" + elif lag_score < -4: + return "明显滞涨" + elif lag_score < -1.5: + return "轻微滞涨" + elif lag_score > 4: + return "跑赢同行" + elif lag_score > 1.5: + return "略胜同行" + else: + return "正常水平" + result = { + 'target_lag_vs_top5': { + 'lag_score': lag_vs_top5, + 'lag_level': get_lag_level(lag_vs_top5), + 'target_change_pct': target_change_pct, + 'top5_avg_change': top5_avg_change + }, + 'target_lag_vs_top1': { + 'lag_score': lag_vs_top1, + 'lag_level': get_lag_level(lag_vs_top1), + 'target_change_pct': target_change_pct, + 'top1_change': top1_change + }, + 'top5_avg_change': top5_avg_change, + 'top1_change': top1_change + } + + logger.info(f"滞涨分析完成 - 目标股票: {target_change_pct:.2f}%, 前5平均: {top5_avg_change:.2f}%, 最高1只: {top1_change:.2f}%") + return result + + except Exception as e: + logger.error(f"滞涨分析失败: {str(e)}") + return { + 'target_lag_vs_top5': None, + 'target_lag_vs_top1': None, + 'top5_avg_change': None, + 'top1_change': None + } + + def get_stock_name(self, stock_code: str) -> Optional[str]: + """ + 获取股票名称 + + Args: + stock_code: 股票代码 + + Returns: + str: 股票名称 + """ + try: + normalized_code = self.normalize_stock_code(stock_code) + + query = text(""" + SELECT gp_name + FROM gp_code_all + WHERE gp_code = :stock_code + LIMIT 1 + """) + + with self.mysql_engine.connect() as conn: + result = conn.execute(query, {"stock_code": normalized_code}).fetchone() + + if result and result[0]: + return result[0] + else: + return None + + except Exception as e: + logger.error(f"获取股票名称失败 {stock_code}: {str(e)}") + return None + + def analyze_stock_overlap(self, stock_code: str, days: int = 20) -> Dict: + """ + 分析股票重叠度的主函数 + + Args: + stock_code: 目标股票代码 + days: 统计涨跌幅的天数 + + Returns: + Dict: 分析结果 + """ + try: + logger.info(f"开始分析股票 {stock_code} 的重叠度") + + # 1. 获取目标股票的行业和概念板块 + industries = self.get_stock_industries(stock_code) + concepts = self.get_stock_concepts(stock_code) + + if not industries and not concepts: + logger.warning(f"股票 {stock_code} 没有找到行业或概念板块数据") + return { + 'target_stock': stock_code, + 'error': '未找到行业或概念板块数据' + } + + # 2. 获取相似股票 + industry_stocks = self.get_similar_stocks_by_industry(industries) + concept_stocks = self.get_similar_stocks_by_concept(concepts) + + # 3. 计算重叠分数 + overlap_scores = self.calculate_overlap_scores(stock_code, industry_stocks, concept_stocks) + + # 4. 筛选高重叠度股票 + high_overlap_stocks = self.filter_high_overlap_stocks(overlap_scores) + + if not high_overlap_stocks: + logger.info(f"股票 {stock_code} 没有找到高重叠度股票（分数>{self.MIN_OVERLAP_SCORE}）") + return { + 'target_stock': stock_code, + 'similar_stocks_count': 0, + 'message': f'没有找到重叠度超过{self.MIN_OVERLAP_SCORE}分的股票' + } + + # 5. 获取前5只重叠度最高的股票 + top5_stocks = high_overlap_stocks[:5] # 取前5只 + top5_stock_codes = [stock[0] for stock in top5_stocks] + all_stock_codes = [self.normalize_stock_code(stock_code)] + top5_stock_codes + + price_changes = self.get_recent_price_changes(all_stock_codes, days) + + # 6. 滞涨分析 + target_change = price_changes.get(self.normalize_stock_code(stock_code)) + similar_changes = {code: data for code, data in price_changes.items() + if code != self.normalize_stock_code(stock_code)} + + lag_analysis = self.analyze_lag_performance(target_change, top5_stocks, similar_changes) + + # 7. 组装前5只股票结果 + top5_results = [] + for similar_stock, overlap_score in top5_stocks: + stock_name = self.get_stock_name(similar_stock) + price_data = price_changes.get(similar_stock, {}) + + top5_results.append({ + 'stock_code': similar_stock, + 'stock_name': stock_name, + 'overlap_score': overlap_score, + 'latest_price': price_data.get('latest_price'), + 'price_change': price_data.get('price_change'), + 'change_pct': price_data.get('change_pct') + }) + + return { + 'target_stock': stock_code, + 'target_stock_name': self.get_stock_name(stock_code), + 'target_price_data': target_change, + 'analysis_date': datetime.now().strftime('%Y-%m-%d %H:%M:%S'), + 'analysis_period_days': days, + 'industries': [{'code': code, 'name': name} for code, name in industries], + 'concepts': [{'code': code, 'name': name} for code, name in concepts], + 'overlap_threshold': self.MIN_OVERLAP_SCORE, + 'top5_stocks': top5_results, + 'lag_analysis': lag_analysis + } + + except Exception as e: + logger.error(f"分析股票重叠度失败 {stock_code}: {str(e)}") + return { + 'target_stock': stock_code, + 'error': str(e) + } + + def close_connection(self): + """关闭数据库连接""" + try: + if self.mysql_engine: + self.mysql_engine.dispose() + logger.info("MySQL连接已关闭") + except Exception as e: + logger.error(f"关闭MySQL连接失败: {str(e)}") + + +def main(): + """主函数 - 示例用法""" + analyzer = None + try: + # 创建分析器实例 + analyzer = OverlapAnalyzer() + + # 示例：分析某只股票的重叠度 + stock_code = "300661.SZ" # 圣邦股份 + + print(f"=== 股票重叠度分析：{stock_code} ===") + + # 执行分析 + result = analyzer.analyze_stock_overlap(stock_code, days=3) + + if 'error' in result: + print(f"分析失败: {result['error']}") + return + + # 输出结果 + print(f"\n目标股票: {result['target_stock_name']} ({result['target_stock']})") + print(f"分析日期: {result['analysis_date']}") + print(f"统计周期: {result['analysis_period_days']}天") + + if 'target_price_data' in result and result['target_price_data']: + target_data = result['target_price_data'] + print(f"目标股票涨跌幅: {target_data.get('change_pct', 0):.2f}%") + + print(f"\n行业板块: {len(result['industries'])}个") + for industry in result['industries'][:3]: # 显示前3个 + print(f" - {industry['name']} ({industry['code']})") + + print(f"\n概念板块: {len(result['concepts'])}个") + for concept in result['concepts'][:3]: # 显示前3个 + print(f" - {concept['name']} ({concept['code']})") + + print(f"\n重叠度最高的5只股票:") + print("-" * 80) + print(f"{'股票代码':<12} {'股票名称':<15} {'重叠分数':<8} {'最新价格':<10} {'涨跌幅':<10}") + print("-" * 80) + + for stock in result['top5_stocks']: + print(f"{stock['stock_code']:<12} " + f"{stock['stock_name'] or '未知':<15} " + f"{stock['overlap_score']:<8.1f} " + f"{stock['latest_price'] or 0:<10.2f} " + f"{stock['change_pct'] or 0:<10.2f}%") + + # 显示滞涨分析结果 + lag_analysis = result.get('lag_analysis', {}) + if lag_analysis: + print(f"\n滞涨分析结果:") + print("-" * 50) + + # vs 前5平均 + vs_top5 = lag_analysis.get('target_lag_vs_top5', {}) + if vs_top5: + print(f"vs 前5平均: {vs_top5.get('lag_score', 0):.2f}% ({vs_top5.get('lag_level', '未知')})") + print(f" 目标股票: {vs_top5.get('target_change_pct', 0):.2f}%") + print(f" 前5平均: {vs_top5.get('top5_avg_change', 0):.2f}%") + + # vs 最高1只 + vs_top1 = lag_analysis.get('target_lag_vs_top1', {}) + if vs_top1: + print(f"vs 最高1只: {vs_top1.get('lag_score', 0):.2f}% ({vs_top1.get('lag_level', '未知')})") + print(f" 目标股票: {vs_top1.get('target_change_pct', 0):.2f}%") + print(f" 最高1只: {vs_top1.get('top1_change', 0):.2f}%") + + except Exception as e: + logger.error(f"程序执行失败: {str(e)}") + finally: + if analyzer: + analyzer.close_connection() + + +if __name__ == "__main__": + main() diff --git a/src/quantitative_analysis/tech_fundamental_factor_strategy.py b/src/quantitative_analysis/tech_fundamental_factor_strategy.py index 7d3a0cc..f36b69d 100644 --- a/src/quantitative_analysis/tech_fundamental_factor_strategy.py +++ b/src/quantitative_analysis/tech_fundamental_factor_strategy.py @@ -713,6 +713,28 @@ def main(): print(f" 最高总分: {df['total_score'].max():.2f}") print(f" 最低总分: {df['total_score'].min():.2f}") + # 生成成长+成熟合并的简表（仅三列：股票代码、总分、排名） + combined_parts = [] + for stage, df in results.items(): + if isinstance(df, pd.DataFrame) and not df.empty: + if 'stock_code' in df.columns and 'total_score' in df.columns: + combined_parts.append(df[['stock_code', 'total_score']].copy()) + + if combined_parts: + combined_df = pd.concat(combined_parts, ignore_index=True) + # 去除总分为空的数据 + combined_df = combined_df.dropna(subset=['total_score']) + # 按总分降序并重新排名 + combined_df = combined_df.sort_values('total_score', ascending=False).reset_index(drop=True) + combined_df['rank'] = range(1, len(combined_df) + 1) + # 保存文件 + combined_file = f"tech_fundamental_factor_all_{datetime.now().strftime('%Y%m%d_%H%M')}.csv" + combined_df.to_csv(combined_file, index=False, encoding='utf-8-sig') + print(f"\n=== 合并结果（成长+成熟） ===") + print(f"总股票数量: {len(combined_df)}") + print(combined_df.head(10).to_string(index=False)) + print(f"\n合并简表已保存到: {combined_file}") + print(f"\n=== 策略运行完成 ===") except Exception as e: diff --git a/src/quantitative_analysis/tech_fundamental_factor_strategy_v2.py b/src/quantitative_analysis/tech_fundamental_factor_strategy_v2.py new file mode 100644 index 0000000..4ed3942 --- /dev/null +++ b/src/quantitative_analysis/tech_fundamental_factor_strategy_v2.py @@ -0,0 +1,757 @@ +#!/usr/bin/env python3 +# -*- coding: utf-8 -*- + +""" +是固态电池、人形机器人、通信设备、算力租赁四个赛道的筛选。 +结合个股历史基本面，以及个股波动情况，这里波动指标是一个波动越大，分数越高的指标 +v2并不是科技主题选股！！！！！！--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 +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 MONGO_CONFIG2, DB_URL + +# 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): + """初始化策略""" + self.lifecycle_calculator = CompanyLifecycleFactor() + self.financial_analyzer = FinancialIndicatorAnalyzer() + 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 + ) + + # 科技概念板块列表 + self.tech_concepts = [ + "固态电池", "人形机器人", "通信设备", "算力租赁" + ] + logger.info("科技主题基本面因子选股策略初始化完成") + + 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}') + UNION + SELECT DISTINCT gp_code as stock_code, gp_name as stock_name, bk_name as concept_name + FROM gp_hybk + WHERE bk_name = '通信设备' + ORDER BY stock_code + """) + + with self.mysql_engine.connect() as conn: + df = pd.read_sql(query, conn) + + logger.info(f"获取到 {len(df)} 只科技概念股票") + 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], year: int = 2024) -> 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 + year: 分析年份 + + Returns: + Dict: 包含成长期和成熟期股票的字典 + """ + try: + logger.info(f"开始分析 {len(stock_codes)} 只股票的企业生命周期") + + # 批量计算生命周期 + lifecycle_df = self.lifecycle_calculator.batch_calculate_lifecycle_factors(stock_codes, 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 + } + + logger.info(f"成长期股票: {len(growth_stage_stocks)} 只") + logger.info(f"成熟期股票: {len(mature_stage_stocks)} 只") + + 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]) -> 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 + + Args: + growth_stocks: 成长期股票列表 + mature_stocks: 成熟期股票列表 + + Returns: + Tuple: (成长期距离因子DataFrame, 成熟期距离因子DataFrame) + """ + try: + growth_distance_df = pd.DataFrame() + mature_distance_df = pd.DataFrame() + + # 计算成长期股票距离因子 + if growth_stocks: + logger.info(f"计算 {len(growth_stocks)} 只成长期股票的距离因子") + growth_data = self.distance_calculator.get_stock_data(growth_stocks) + if not growth_data.empty: + growth_indicators = self.distance_calculator.calculate_technical_indicators(growth_data) + 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: + logger.info(f"计算 {len(mature_stocks)} 只成熟期股票的距离因子") + mature_data = self.distance_calculator.get_stock_data(mature_stocks) + if not mature_data.empty: + mature_indicators = self.distance_calculator.calculate_technical_indicators(mature_data) + 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)}") + 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: + logger.info(f"计算 {len(stock_codes)} 只股票的通用因子") + + results = [] + latest_date = "2025-06-30" # 最新季度数据 + annual_date = "2024-12-31" # 年报数据 + + for stock_code in stock_codes: + try: + + factor_data = {'stock_code': stock_code} + + # 1. 毛利率（使用最新数据） + gross_margin = self.financial_analyzer.analyze_gross_profit_margin(stock_code, latest_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, 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, 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) + logger.info(f"成功计算 {len(df)} 只股票的通用因子") + 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: + logger.info(f"计算 {len(stock_codes)} 只成长期股票的特色因子") + + results = [] + latest_date = "2025-06-30" # 使用最新数据 + annual_date = "2024-12-31" # 使用年度数据 + + for stock_code in stock_codes: + try: + + factor_data = {'stock_code': stock_code} + + # 1. 管理费用率（使用最新数据） + admin_ratio = self.financial_analyzer.analyze_admin_expense_ratio(stock_code, latest_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, latest_date) + financial_data = self.financial_analyzer.get_financial_data(stock_code, 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（使用最新数据） + asset_liability_ratio = self.financial_analyzer.analyze_asset_liability_ratio(stock_code, latest_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) + logger.info(f"成功计算 {len(df)} 只成长期股票的特色因子") + 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: + logger.info(f"计算 {len(stock_codes)} 只成熟期股票的特色因子") + + latest_date = "2025-06-30" # 使用最新数据 + + # 在循环外获取全A股PB和ROE数据，避免重复查询 + logger.info("获取全A股PB数据...") + all_pb_data = self.financial_analyzer.get_all_stocks_pb_data() + + logger.info("获取全A股ROE数据...") + all_roe_data = self.financial_analyzer.get_all_stocks_roe_data(latest_date) + + results = [] + + for stock_code in stock_codes: + try: + factor_data = {'stock_code': stock_code} + + # 1. 应收账款周转率（使用最新数据） + formatted_stock_code = self.financial_analyzer.code_formatter.to_dot_format(stock_code) + financial_data = self.financial_analyzer.get_financial_data(formatted_stock_code, latest_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（使用最新数据） + rd_intensity = self.financial_analyzer.analyze_rd_expense_ratio(stock_code, latest_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) + logger.info(f"成功计算 {len(df)} 只成熟期股票的特色因子") + 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, year: int = 2024) -> Dict[str, pd.DataFrame]: + """ + 运行完整的选股策略 + + Args: + year: 分析年份 + + Returns: + Dict: 包含成长期和成熟期股票分析结果的字典 + """ + try: + logger.info("开始运行科技主题基本面因子选股策略") + + # 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() + logger.info(f"共获取到 {len(stock_codes)} 只科技概念股票") + + # 2. 按企业生命周期筛选 + lifecycle_result = self.filter_by_lifecycle(stock_codes, year) + 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 + logger.info(f"成长期结果: {len(growth_result)} 只股票") + + # 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 + logger.info(f"成熟期结果: {len(mature_result)} 只股票") + + logger.info("科技主题基本面因子选股策略运行完成") + 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': True, # 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] + + logger.info(f"完成 {stage} 阶段 {len(df_scored)} 只股票的因子打分") + 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': 2/8, + + # 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': 2/8, + + # 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) + + logger.info(f"完成 {stage} 阶段 {len(df_result)} 只股票的总分计算") + 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 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() + logger.info("数据库连接已关闭") + 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: + print("=== 科技主题基本面因子选股策略 ===") + print("数据说明:") + print("- 毛利率、净利润增长率等：使用最新数据 (2025-06-30)") + print("- 供应商客户集中度、折旧摊销、研发费用：使用年报数据 (2024-12-31)") + print() + + # 创建策略实例 + strategy = TechFundamentalFactorStrategy() + logger.info("策略实例创建成功") + + # 运行策略 + results = strategy.run_strategy(year=2024) + + # 输出结果 + if not results: + print("未获得分析结果") + return + + for stage, df in results.items(): + print(f"\n=== {stage.upper()} 阶段股票分析结果 ===") + print(f"股票数量: {len(df)}") + + if not df.empty: + # 调试：显示所有列名 + print(f"数据列: {list(df.columns)}") + # 显示前5只股票的关键指标 + print("\n前5只股票:") + display_columns = [ + 'stock_code', 'gross_profit_margin', 'growth_score', + 'supplier_concentration', 'customer_concentration', + 'total_score', 'rank' + ] + available_columns = [col for col in display_columns if col in df.columns] + print(df[available_columns].head(5).to_string(index=False)) + + # 保存完整结果 + output_file = f"tech_fundamental_factor_{stage}_{datetime.now().strftime('%Y%m%d_%H%M')}.csv" + df.to_csv(output_file, index=False, encoding='utf-8-sig') + print(f"\n完整结果已保存到: {output_file}") + + # 显示统计信息 + print(f"\n统计信息:") + print(f" 平均总分: {df['total_score'].mean():.2f}") + print(f" 最高总分: {df['total_score'].max():.2f}") + print(f" 最低总分: {df['total_score'].min():.2f}") + + # 生成成长+成熟合并的简表（仅三列：股票代码、总分、排名） + combined_parts = [] + for stage, df in results.items(): + if isinstance(df, pd.DataFrame) and not df.empty: + if 'stock_code' in df.columns and 'total_score' in df.columns: + combined_parts.append(df[['stock_code', 'total_score']].copy()) + + if combined_parts: + combined_df = pd.concat(combined_parts, ignore_index=True) + # 去除总分为空的数据 + combined_df = combined_df.dropna(subset=['total_score']) + # 按总分降序并重新排名 + combined_df = combined_df.sort_values('total_score', ascending=False).reset_index(drop=True) + combined_df['rank'] = range(1, len(combined_df) + 1) + # 追加概念列：获取每个股票所属概念（若多概念则以逗号分隔） + try: + concept_df = strategy.get_tech_stocks() + if not concept_df.empty and 'stock_code' in concept_df.columns and 'concept_name' in concept_df.columns: + concept_map = ( + concept_df.groupby('stock_code')['concept_name'] + .apply(lambda s: ','.join(sorted(set([str(x) for x in s if pd.notna(x)])))) + .to_dict() + ) + combined_df['concepts'] = combined_df['stock_code'].map(concept_map) + else: + combined_df['concepts'] = None + except Exception: + combined_df['concepts'] = None + # 保存文件 + combined_file = f"tech_fundamental_factor_all_{datetime.now().strftime('%Y%m%d_%H%M')}.csv" + combined_df.to_csv(combined_file, index=False, encoding='utf-8-sig') + print(f"\n=== 合并结果（成长+成熟） ===") + print(f"总股票数量: {len(combined_df)}") + print(combined_df.head(10).to_string(index=False)) + print(f"\n合并简表已保存到: {combined_file}") + + print(f"\n=== 策略运行完成 ===") + + 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() \ No newline at end of file diff --git a/src/quantitative_analysis/us_batch_stock_price_collector.py b/src/quantitative_analysis/us_batch_stock_price_collector.py index 576f549..8379b9c 100644 --- a/src/quantitative_analysis/us_batch_stock_price_collector.py +++ b/src/quantitative_analysis/us_batch_stock_price_collector.py @@ -1,3 +1,5 @@ +#美股行业板块导出，从通达信里面打开美股首页，然后栏目里面有个细分行业。点击导出所有栏目即可 + import requests import pandas as pd from datetime import datetime @@ -291,3 +293,5 @@ if __name__ == '__main__': fetch_and_store_us_stock_data_optimized(use_proxy=False) # 默认不使用代理 + + diff --git a/src/scripts/config.py b/src/scripts/config.py index ce9ed01..400fc39 100644 --- a/src/scripts/config.py +++ b/src/scripts/config.py @@ -11,7 +11,7 @@ XUEQIU_HEADERS = { 'Accept-Encoding': 'gzip, deflate, br, zstd', 'Accept-Language': 'zh-CN,zh;q=0.9,en;q=0.8', 'Client-Version': 'v2.44.75', - 'Cookie': 'cookiesu=811743062689927; device_id=33fa3c7fca4a65f8f4354e10ed6b7470; smidV2=20250327160437f244626e8b47ca2a7992f30f389e4e790074ae48656a22f10; HMACCOUNT=8B64A2E3C307C8C0; s=c611ttmqlj; xq_is_login=1; u=8493411634; bid=4065a77ca57a69c83405d6e591ab5449_m8r2nhs8; __utma=1.434320573.1747189698.1747189698.1747189698.1; __utmc=1; __utmz=1.1747189698.1.1.utmcsr=(direct)|utmccn=(direct)|utmcmd=(none); snbim_minify=true; _c_WBKFRo=dsWgHR8i8KGPbIyhFlN51PHOzVuuNytvUAFppfkD; _nb_ioWEgULi=; aliyungf_tc=00c6b999835b16cea9e4a6aab36cca373a0976bf55ee74770d11f421f7119ad8; Hm_lvt_1db88642e346389874251b5a1eded6e3=1757464900; xq_a_token=975ab9f15a4965b9e557b9bc6f33bc1da20a0f49; xqat=975ab9f15a4965b9e557b9bc6f33bc1da20a0f49; xq_id_token=eyJ0eXAiOiJKV1QiLCJhbGciOiJSUzI1NiJ9.eyJ1aWQiOjg0OTM0MTE2MzQsImlzcyI6InVjIiwiZXhwIjoxNzYwMjU2NzM4LCJjdG0iOjE3NTc2NjQ3Mzg4MjUsImNpZCI6ImQ5ZDBuNEFadXAifQ.TMx4-TjKx96j5h6-EGiRIM2WKtJm1xctZhYidc40Em0pRcr0UBHAKBGl3No5r1BElYa9qnEDgNYI0Zv137Inx-EMPqm5cd1Z_ZjLdWOSLzT9qqBj8zdfuqJwP2nCYvC6KLjd8BvykS0vSFKqwb-r0WhEA3OzbO8teVNsaemdKAhBoIyP3-RQCfRxJ9RLNha1ZMdg66iZvfz_SOsG41y8IA9yyl-FFFJOq4TnAiywY1yO1QIJJhkh8YQqfnDfQQdSIFgJGToU980Lw1dm4aCDY-kvn-t18KjrL_hZJ_UNN65bgZsSsuWf-VQ7wsjjczNrfBYAHdZ6kES0CGo9g8IZZw; xq_r_token=c209224335327f29fc555d9910b43c0df6d52d5a; Hm_lpvt_1db88642e346389874251b5a1eded6e3=1757774901; .thumbcache_f24b8bbe5a5934237bbc0eda20c1b6e7=rJD0qKtipTMjRBkLBVEyXbl0CiVeY7y4AxEZC0Vf6Zkou9cxp0NPsxwSrnOyFyBMr+Ws5/nJDO1NUalRDyAPsA%3D%3D; acw_tc=3ccdc14717578973700474899e2dc1c35b6358f1af81617250f8f00b4cf31c; ssxmod_itna=1-eqGxBDnGKYuxcD4kDRgxYq7ueYKS8DBP01Dp2xQyP08D60DB40Q0qRHhqDylAkiETFN1t42Y5D/KlYeDZDGFdDqx0Ei6FiYHK1ezjCGbKSAQY5P53Niio89NQ7DEQm6fjL1S4K7s5h8KRDo9n4hiDB3DbqDymgY5qxGGA4GwDGoD34DiDDPDb8rDALeD7qDFnenropTDm4GWneGfDDoDYbT3xiUYDDUvbeG2iET4DDN4bIGYZ2G76=r1doBip29xKiTDjqPD/ShUoiuzZKC4icFL2/amAeGyC5GuY6mWHQ77SWcbscAV70i8hx_Bx3rKqB5YGDRYqK8o2xY9iKR0YRDxeEDW0DWnQ8EwhDDiP46iRiGDYZtgrNMhXiY4MQA7bAilP4nPkFGCmqzBqGYesQGQiT3ihKbm5CexbxxD; ssxmod_itna2=1-eqGxBDnGKYuxcD4kDRgxYq7ueYKS8DBP01Dp2xQyP08D60DB40Q0qRHhqDylAkiETFN1t42YeDA4rYnRItORCitz/D3nyGQigbiD', + 'Cookie': 'cookiesu=811743062689927; device_id=33fa3c7fca4a65f8f4354e10ed6b7470; smidV2=20250327160437f244626e8b47ca2a7992f30f389e4e790074ae48656a22f10; HMACCOUNT=8B64A2E3C307C8C0; s=c611ttmqlj; xq_is_login=1; u=8493411634; bid=4065a77ca57a69c83405d6e591ab5449_m8r2nhs8; __utma=1.434320573.1747189698.1747189698.1747189698.1; __utmc=1; __utmz=1.1747189698.1.1.utmcsr=(direct)|utmccn=(direct)|utmcmd=(none); snbim_minify=true; _c_WBKFRo=dsWgHR8i8KGPbIyhFlN51PHOzVuuNytvUAFppfkD; _nb_ioWEgULi=; aliyungf_tc=00c6b999835b16cea9e4a6aab36cca373a0976bf55ee74770d11f421f7119ad8; Hm_lvt_1db88642e346389874251b5a1eded6e3=1757464900; xq_a_token=0c9cfffc3d713428b4759d3c4e11fff91b618525; xqat=0c9cfffc3d713428b4759d3c4e11fff91b618525; xq_id_token=eyJ0eXAiOiJKV1QiLCJhbGciOiJSUzI1NiJ9.eyJ1aWQiOjg0OTM0MTE2MzQsImlzcyI6InVjIiwiZXhwIjoxNzYxNTYzMTU1LCJjdG0iOjE3NTg5NzExNTUyODksImNpZCI6ImQ5ZDBuNEFadXAifQ.KotT161D0bQF-m3HNgQFTknHqUpMUR_KWY7BAjiLKOaQocVghM85sCKUretUK5OpMDXMzGPhzxTWMs4VhGxKUHxHSu3inrbCw7rgGDtK0Lnr_6OXHkEMwDadN37V3AYbxKWaX609pACLK1q2cvRmGg94wAgj9KEqu76gdV4w4tRv4LupUJQgYb6M0GEVe7ejLs2_CdV5PV7j0PSjTtpjaeZWkc9YhW32vOyA6gg3fU9S4BQg7qgciYLI8AxiU-GYgFLyFSK2S7nvAngz8TEhNoQcClKrD4NeqqSj2nA5bh5_5FM8AbdbKGw6K8eIO0wkxCV2iUWBDMz-h160cX0GLw; xq_r_token=e586d0abb8de83a22a5d66be2f908954a50005b2; .thumbcache_f24b8bbe5a5934237bbc0eda20c1b6e7=nGvY6/KwbtfD3HgPBzY+S0khnrVCqr4mPyfsZ27X+GjYcRWS61PMogABu5uvj/LAd4TJsSFQiYwdEepZMxwg3A%3D%3D; ssxmod_itna=1-iqmxyQG=KWqGqwD4qeqY5Lor=G2zDuDBP01DpxYK0CDmxjKideDUDtTxR3R27rQA04D7777mqqeCQ_4D/A0ieDZDGKQDqx0or0KBiBUPehSmRxdW9gTBThssUUo8vG8FV4x9WCL5ZG9emiTjxIx_D0aDmKDUx4x_BAYDeWFDCeDQxirDD4DAmPDFxibDimAb4DdkK8EvUOpIDGrDlKDRchok4GWDiPD76gDl6gpbDDCDiyjKV3Dixiaz3eDBomc41iI4xi3kb8_c8b5Yq4EWi2ID7v3DlPqjdnIKy/M4E3Eps/9=agju40ODG6CofPchOBoFO3o6=rlwIlqrDKpY4Kleiix57xhx_eQDrWD57IpDx54_xmmZ7hpY21Bb0nQi/AO4b3mDrnNDt0=a0q345Bi5Drhl54nmeYnAD_mmD_GWq7h1iGq7YzGxA8DqBD=4nz7xNImYnb47xxD; ssxmod_itna2=1-iqmxyQG=KWqGqwD4qeqY5Lor=G2zDuDBP01DpxYK0CDmxjKideDUDtTxR3R27rQA04D7777mqqeCQ4DW_QRiAwxKA07vi3WDj4pxkh3pfaDlxTMHwkxxRl_UFFyKFmUR1v7kCiqQHALxyxuseYn8Q79bLg08eBKxkZwGFEy8ZxTGW7YkIrrzM=5YCx_Gem5Qg2HvQz9AhR0FvxfOG7AQwzoPjSpGmtApnKdwC4uUuzI6eDo2Bedxmf8L2fgms1KnurZO0K=m0ifvg8yDYb=qs=GimaBFojOsWf01ibDy6ZP5DLpur3mS//PdQPR0PqEEqiKseLxQQ3iKiC3ijQ5zKLELcAULPdKD6Q86ejjQKAk6K3leiePYGxdGpYpDbO3YPKa_x4RPTxTif5T3mOIGa8TCY44Y4hWn1AQq7G1nDQ7GQ7D5BDxzWw8bz3D; is_overseas=0; Hm_lpvt_1db88642e346389874251b5a1eded6e3=1759979780', 'Referer': 'https://weibo.com/u/7735765253', 'Sec-Ch-Ua': '"Chromium";v="122", "Not(A:Brand";v="24", "Google Chrome";v="122"', 'Sec-Ch-Ua-Mobile': '?0',