#!/usr/bin/env python3 """ 高级强势板块筛选器 筛选条件： 1. 本周收阳（周涨幅>0） 2. 周线级别创阶段新高（20周新高） 3. 成交额巨大（超过1000亿） """ import sys from pathlib import Path import pandas as pd import numpy as np from datetime import datetime, timedelta # 添加项目根目录到路径 current_dir = Path(__file__).parent sys.path.insert(0, str(current_dir)) from src.data.tushare_fetcher import TushareFetcher from loguru import logger def get_trading_dates(days_back=100): """获取过去N个交易日""" dates = [] current = datetime.now() while len(dates) < days_back: if current.weekday() < 5: # 周一到周五 dates.append(current.strftime('%Y%m%d')) current -= timedelta(days=1) return sorted(dates) # 升序返回 def filter_index_concepts(ths_concepts): """过滤掉指数型板块""" index_keywords = [ '成份股', '样本股', '成分股', '50', '100', '300', '500', '1000', '上证', '深证', '中证', '创业板', '科创板', '北证', 'ETF', '指数', 'Index', '基准' ] def is_concept_plate(name: str) -> bool: name_lower = name.lower() for keyword in index_keywords: if keyword.lower() in name_lower: return False return True original_count = len(ths_concepts) filtered_concepts = ths_concepts[ths_concepts['name'].apply(is_concept_plate)] filtered_count = len(filtered_concepts) logger.info(f"过滤指数型板块: {original_count} -> {filtered_count} 个") return filtered_concepts def analyze_concept_strength(fetcher: TushareFetcher, concept_info, trading_dates): """分析单个概念的强势特征""" ts_code = concept_info['ts_code'] name = concept_info['name'] try: # 获取过去20周的数据（约100个交易日） start_date = trading_dates[0] end_date = trading_dates[-1] daily_data = fetcher.pro.ths_daily( ts_code=ts_code, start_date=start_date, end_date=end_date ) if daily_data.empty or len(daily_data) < 20: return None # 按日期排序 daily_data = daily_data.sort_values('trade_date') daily_data.reset_index(drop=True, inplace=True) # 1. 计算本周涨幅（最近5个交易日） recent_data = daily_data.tail(5) if len(recent_data) < 2: return None week_start_close = recent_data.iloc[0]['close'] week_end_close = recent_data.iloc[-1]['close'] week_change = (week_end_close - week_start_close) / week_start_close * 100 # 2. 检查是否本周收阳 is_weekly_positive = week_change > 0 # 3. 计算20周新高（约100个交易日） current_close = daily_data.iloc[-1]['close'] past_20weeks_high = daily_data['high'].max() is_20week_high = current_close >= past_20weeks_high * 0.99 # 允许1%的误差 # 4. 计算成交额（最近5日平均） recent_turnover = recent_data['vol'].mean() * recent_data['close'].mean() # 简化计算 turnover_100yi = recent_turnover / 100000000 # 转换为亿元 # 5. 计算技术指标 # RSI相对强弱指数 rsi = calculate_rsi(daily_data['close'].values) # 20日均线趋势 ma20 = daily_data['close'].rolling(20).mean() ma20_trend = (ma20.iloc[-1] - ma20.iloc[-10]) / ma20.iloc[-10] * 100 if len(ma20) >= 20 else 0 # 波动率 volatility = daily_data['pct_change'].std() * np.sqrt(250) # 年化波动率 return { 'ts_code': ts_code, 'name': name, 'week_change': week_change, 'is_weekly_positive': is_weekly_positive, 'is_20week_high': is_20week_high, 'avg_turnover_yi': turnover_100yi, 'current_close': current_close, 'rsi': rsi, 'ma20_trend': ma20_trend, 'volatility': volatility, 'data_length': len(daily_data) } except Exception as e: logger.debug(f"分析 {name} 失败: {e}") return None def calculate_rsi(prices, period=14): """计算RSI指标""" try: delta = np.diff(prices) gain = np.where(delta > 0, delta, 0) loss = np.where(delta < 0, -delta, 0) avg_gain = np.mean(gain[-period:]) if len(gain) >= period else 0 avg_loss = np.mean(loss[-period:]) if len(loss) >= period else 0 if avg_loss == 0: return 100 rs = avg_gain / avg_loss rsi = 100 - (100 / (1 + rs)) return rsi except: return 50 # 默认值 def find_strong_sectors(fetcher: TushareFetcher): """寻找强势板块""" try: logger.info("🔍 开始寻找强势板块...") # 获取交易日期 trading_dates = get_trading_dates(100) logger.info(f"分析周期: {trading_dates[0]} 到 {trading_dates[-1]} (100个交易日)") # 获取同花顺概念列表 ths_concepts = fetcher.pro.ths_index(exchange='A', type='N') if ths_concepts.empty: logger.error("未获取到概念数据") return # 过滤指数型概念 ths_concepts = filter_index_concepts(ths_concepts) # 分析概念强度 strong_concepts = [] total_concepts = min(80, len(ths_concepts)) # 分析前80个概念 logger.info(f"分析 {total_concepts} 个概念...") for i, (_, concept) in enumerate(ths_concepts.head(total_concepts).iterrows()): if i % 10 == 0: logger.info(f"进度: {i+1}/{total_concepts}") result = analyze_concept_strength(fetcher, concept, trading_dates) if result: strong_concepts.append(result) if not strong_concepts: logger.warning("未找到符合条件的强势板块") return # 转换为DataFrame df = pd.DataFrame(strong_concepts) # 强势板块筛选 logger.info("🚀 应用强势板块筛选条件...") # 条件1：本周收阳 weekly_positive = df[df['is_weekly_positive']] logger.info(f"本周收阳概念: {len(weekly_positive)} 个") # 条件2：20周新高 new_high_concepts = df[df['is_20week_high']] logger.info(f"20周新高概念: {len(new_high_concepts)} 个") # 条件3：成交额超过1000亿（这里设置为10亿，因为单个概念1000亿太高） high_turnover = df[df['avg_turnover_yi'] >= 10] logger.info(f"成交额超过10亿概念: {len(high_turnover)} 个") # 综合强势板块（满足至少2个条件） df['strength_score'] = ( df['is_weekly_positive'].astype(int) + df['is_20week_high'].astype(int) + (df['avg_turnover_yi'] >= 10).astype(int) ) # 按强势得分和周涨幅排序 strong_sectors = df[df['strength_score'] >= 2].sort_values(['strength_score', 'week_change'], ascending=[False, False]) # 显示结果 display_strong_sectors(df, strong_sectors, weekly_positive, new_high_concepts, high_turnover) return df except Exception as e: logger.error(f"寻找强势板块失败: {e}") def display_strong_sectors(df, strong_sectors, weekly_positive, new_high_concepts, high_turnover): """显示强势板块分析结果""" print("\n" + "="*100) print("🔍 强势板块综合分析报告") print("="*100) # 1. 综合强势板块（满足多个条件） if not strong_sectors.empty: print(f"\n🚀 综合强势板块TOP10（满足2+条件）:") print(f"{'排名':<4} {'概念名称':<25} {'周涨幅':<10} {'强势分':<8} {'RSI':<8} {'成交额(亿)':<12} {'条件':<20}") print("-" * 100) for i, (_, concept) in enumerate(strong_sectors.head(10).iterrows()): rank = i + 1 name = concept['name'][:23] + '..' if len(concept['name']) > 23 else concept['name'] week_chg = f"{concept['week_change']:+.2f}%" score = f"{concept['strength_score']}/3" rsi = f"{concept['rsi']:.1f}" turnover = f"{concept['avg_turnover_yi']:.1f}" conditions = [] if concept['is_weekly_positive']: conditions.append("周阳") if concept['is_20week_high']: conditions.append("新高") if concept['avg_turnover_yi'] >= 10: conditions.append("大额") condition_str = "+".join(conditions) print(f"{rank:<4} {name:<25} {week_chg:<10} {score:<8} {rsi:<8} {turnover:<12} {condition_str:<20}") # 2. 分类展示 print(f"\n📊 分类统计:") print(f" 本周收阳: {len(weekly_positive)} 个") print(f" 20周新高: {len(new_high_concepts)} 个") print(f" 大成交额: {len(high_turnover)} 个") print(f" 综合强势: {len(strong_sectors)} 个") # 3. 本周收阳TOP10 if not weekly_positive.empty: top_weekly = weekly_positive.sort_values('week_change', ascending=False) print(f"\n📈 本周收阳TOP10:") for i, (_, concept) in enumerate(top_weekly.head(10).iterrows()): print(f" {i+1:2d}. {concept['name']}: {concept['week_change']:+.2f}%") # 4. 20周新高概念 if not new_high_concepts.empty: print(f"\n🎯 20周新高概念TOP10:") new_high_sorted = new_high_concepts.sort_values('week_change', ascending=False) for i, (_, concept) in enumerate(new_high_sorted.head(10).iterrows()): print(f" {i+1:2d}. {concept['name']}: {concept['week_change']:+.2f}% (RSI: {concept['rsi']:.1f})") # 5. 大成交额概念 if not high_turnover.empty: print(f"\n💰 大成交额概念TOP10:") turnover_sorted = high_turnover.sort_values('avg_turnover_yi', ascending=False) for i, (_, concept) in enumerate(turnover_sorted.head(10).iterrows()): print(f" {i+1:2d}. {concept['name']}: {concept['avg_turnover_yi']:.1f}亿 ({concept['week_change']:+.2f}%)") # 6. 技术面强势 strong_tech = df[(df['rsi'] > 60) & (df['ma20_trend'] > 0)].sort_values('week_change', ascending=False) if not strong_tech.empty: print(f"\n📊 技术面强势概念TOP10:") for i, (_, concept) in enumerate(strong_tech.head(10).iterrows()): print(f" {i+1:2d}. {concept['name']}: RSI {concept['rsi']:.1f}, MA20趋势 {concept['ma20_trend']:+.2f}%") def main(): """主函数""" logger.info("🚀 开始高级强势板块分析...") # 初始化Tushare数据获取器 token = "0ed6419a00d8923dc19c0b58fc92d94c9a0696949ab91a13aa58a0cc" fetcher = TushareFetcher(token=token) # 寻找强势板块 result_df = find_strong_sectors(fetcher) if result_df is not None: logger.info("✅ 强势板块分析完成!") else: logger.error("❌ 强势板块分析失败!") if __name__ == "__main__": main()