|
|
@@ -0,0 +1,1783 @@
|
|
|
+"""
|
|
|
+期货市场技术形态规律分析 - 综合版本
|
|
|
+研究期货品种在大幅下跌后出现特定K线形态的规律，并分析后续走势
|
|
|
+
|
|
|
+本文件包含：
|
|
|
+- 分析配置参数
|
|
|
+- 主要分析类FuturePatternAnalyzer
|
|
|
+- 基于合约切换日志的周线数据获取功能
|
|
|
+- 辅助工具函数和数据导出功能
|
|
|
+- 完整的分析流程
|
|
|
+
|
|
|
+主要功能：
|
|
|
+1. get_contract_weekly_data() - 基于合约切换日志获取周线数据的正式方法
|
|
|
+2. get_weekly_data_by_contract_log() - 核心数据获取逻辑，直接使用get_bars获取原生周线数据
|
|
|
+3. 简化的get_weekly_kline_data() - 直接使用8888加权合约，避免主力合约切换复杂性
|
|
|
+4. validate_manual_patterns() - 手动模式验证系统，支持指定日期的模式验证
|
|
|
+5. run_manual_pattern_validation() - 便捷的手动验证入口函数
|
|
|
+6. _display_full_candlestick_charts() - 模式识别失败时的完整K线图显示
|
|
|
+
|
|
|
+数据结构简化：
|
|
|
+- 移除了主力合约获取和切换逻辑
|
|
|
+- 使用单级字典结构：键为8888合约代码，值为日期范围
|
|
|
+- 所有数据检索均使用8888加权合约，简化了代码复杂性
|
|
|
+
|
|
|
+验证模式：
|
|
|
+- 自动搜索模式：使用identify_target_pattern()自动搜索符合条件的技术形态
|
|
|
+- 手动验证模式：使用validate_manual_patterns()验证指定日期是否符合模式条件
|
|
|
+- 两种模式使用相同的验证逻辑和标准，但工作流程独立
|
|
|
+
|
|
|
+作者: jukuan研究团队
|
|
|
+日期: 2025-09
|
|
|
+适用平台: 聚宽在线研究平台
|
|
|
+"""
|
|
|
+
|
|
|
+import pandas as pd
|
|
|
+import numpy as np
|
|
|
+from jqdata import *
|
|
|
+import datetime
|
|
|
+import matplotlib.pyplot as plt
|
|
|
+import matplotlib.patches as patches
|
|
|
+from matplotlib.dates import DateFormatter
|
|
|
+import warnings
|
|
|
+warnings.filterwarnings('ignore')
|
|
|
+
|
|
|
+# 中文字体设置
|
|
|
+plt.rcParams['font.sans-serif'] = ['SimHei', 'Microsoft YaHei', 'DejaVu Sans']
|
|
|
+plt.rcParams['axes.unicode_minus'] = False
|
|
|
+
|
|
|
+# =====================================================================================
|
|
|
+# 第一部分：分析配置参数
|
|
|
+# 用户可根据需要调整以下参数来控制分析行为
|
|
|
+# =====================================================================================
|
|
|
+
|
|
|
+class AnalysisConfig:
|
|
|
+ """分析配置类 - 集中管理所有分析参数"""
|
|
|
+
|
|
|
+ # ==================== 分析时间设置 ====================
|
|
|
+ START_DATE = '2022-01-01' # 分析起始日期 (建议至少2年数据)
|
|
|
+ END_DATE = '2025-08-31' # 分析结束日期
|
|
|
+
|
|
|
+ # ==================== 筛选条件设置 ====================
|
|
|
+ MIN_DRAWDOWN_THRESHOLD = 0.30 # 最小回撤阈值 (0.3 = 30%)
|
|
|
+
|
|
|
+ # K线形态识别条件
|
|
|
+ KLINE_CONDITIONS = {
|
|
|
+ 'max_open_close_diff': 0.01, # 开盘价与收盘价最大差异 (0.01 = 1%)
|
|
|
+ 'min_high_open_gain': 0.03, # 最高价相对开盘价的最小涨幅 (0.03 = 3%)
|
|
|
+ }
|
|
|
+
|
|
|
+ # ==================== 分析参数设置 ====================
|
|
|
+ ANALYSIS_EXTEND_DAYS = 30 # 扩展分析天数 (在下跌期间前后各扩展的天数)
|
|
|
+ FOLLOW_UP_WEEKS = 10 # 后续跟踪周数
|
|
|
+ SUCCESS_RETURN_THRESHOLD = 0.05 # 成功率计算的收益率阈值 (0.05 = 5%)
|
|
|
+
|
|
|
+ # ==================== 输出设置 ====================
|
|
|
+ GENERATE_CHARTS = True # 是否生成K线图
|
|
|
+ MAX_CHART_COUNT = 10 # 最大图表生成数量 (避免生成过多图片)
|
|
|
+ CHART_OUTPUT_DIR = './' # 图表保存目录
|
|
|
+ CSV_ENCODING = 'utf-8-sig' # CSV文件编码
|
|
|
+
|
|
|
+ # ==================== 期货品种过滤设置 ====================
|
|
|
+ # 需要排除的期货品种 (可根据需要调整)
|
|
|
+ EXCLUDED_FUTURES = [
|
|
|
+ 'IF9999.CCFX', #
|
|
|
+ 'LU8888.XINE', #
|
|
|
+ 'NR8888.XINE', #
|
|
|
+ 'LG8888.XDCE', # 原木 - 新上市品种数据较少
|
|
|
+ ]
|
|
|
+
|
|
|
+ # 重点关注的期货品种 (如果设置，将只分析这些品种)
|
|
|
+ FOCUS_FUTURES = [] # 空列表表示分析所有品种
|
|
|
+
|
|
|
+ # ==================== 分析规模控制 ====================
|
|
|
+ MAX_ANALYSIS_FUTURES = 5 # 最大分析期货品种数 (-1表示分析所有品种)
|
|
|
+ VERBOSE_LOGGING = True # 是否打印详细日志
|
|
|
+ SHOW_PROGRESS = True # 是否在分析过程中显示进度
|
|
|
+
|
|
|
+ # ==================== 数据质量控制 ====================
|
|
|
+ MIN_DATA_LENGTH = 100 # 最小数据长度要求 (日线数据点数)
|
|
|
+ MIN_WEEKLY_DATA = 20 # 最小周线数据要求
|
|
|
+ MAX_MISSING_DATA_RATIO = 0.1 # 最大缺失数据比例
|
|
|
+
|
|
|
+ @classmethod
|
|
|
+ def get_config_dict(cls):
|
|
|
+ """获取分析配置字典"""
|
|
|
+ return {
|
|
|
+ 'time_range': {
|
|
|
+ 'start_date': cls.START_DATE,
|
|
|
+ 'end_date': cls.END_DATE,
|
|
|
+ },
|
|
|
+ 'filter_conditions': {
|
|
|
+ 'min_drawdown': cls.MIN_DRAWDOWN_THRESHOLD,
|
|
|
+ 'kline_conditions': cls.KLINE_CONDITIONS,
|
|
|
+ },
|
|
|
+ 'analysis_params': {
|
|
|
+ 'extend_days': cls.ANALYSIS_EXTEND_DAYS,
|
|
|
+ 'follow_up_weeks': cls.FOLLOW_UP_WEEKS,
|
|
|
+ 'success_threshold': cls.SUCCESS_RETURN_THRESHOLD,
|
|
|
+ },
|
|
|
+ 'output_settings': {
|
|
|
+ 'generate_charts': cls.GENERATE_CHARTS,
|
|
|
+ 'max_chart_count': cls.MAX_CHART_COUNT,
|
|
|
+ 'chart_dir': cls.CHART_OUTPUT_DIR,
|
|
|
+ 'csv_encoding': cls.CSV_ENCODING,
|
|
|
+ },
|
|
|
+ 'future_filter': {
|
|
|
+ 'excluded': cls.EXCLUDED_FUTURES,
|
|
|
+ 'focus_only': cls.FOCUS_FUTURES,
|
|
|
+ },
|
|
|
+ 'analysis_control': {
|
|
|
+ 'max_futures': cls.MAX_ANALYSIS_FUTURES,
|
|
|
+ 'verbose': cls.VERBOSE_LOGGING,
|
|
|
+ 'show_progress': cls.SHOW_PROGRESS,
|
|
|
+ },
|
|
|
+ 'data_quality': {
|
|
|
+ 'min_data_length': cls.MIN_DATA_LENGTH,
|
|
|
+ 'min_weekly_data': cls.MIN_WEEKLY_DATA,
|
|
|
+ 'max_missing_ratio': cls.MAX_MISSING_DATA_RATIO,
|
|
|
+ }
|
|
|
+ }
|
|
|
+
|
|
|
+ @classmethod
|
|
|
+ def print_config(cls):
|
|
|
+ """打印当前配置信息"""
|
|
|
+ print("=== 期货技术形态分析配置 ===")
|
|
|
+ print(f"分析时间范围: {cls.START_DATE} 至 {cls.END_DATE}")
|
|
|
+ print(f"最小回撤阈值: {cls.MIN_DRAWDOWN_THRESHOLD*100:.0f}%")
|
|
|
+ print(f"开收差异阈值: {cls.KLINE_CONDITIONS['max_open_close_diff']*100:.0f}%")
|
|
|
+ print(f"最高涨幅阈值: {cls.KLINE_CONDITIONS['min_high_open_gain']*100:.0f}%")
|
|
|
+ print(f"后续跟踪周数: {cls.FOLLOW_UP_WEEKS}")
|
|
|
+ print(f"成功率阈值: {cls.SUCCESS_RETURN_THRESHOLD*100:.0f}%")
|
|
|
+
|
|
|
+ if cls.EXCLUDED_FUTURES:
|
|
|
+ print(f"排除品种数: {len(cls.EXCLUDED_FUTURES)}")
|
|
|
+ if cls.FOCUS_FUTURES:
|
|
|
+ print(f"重点品种数: {len(cls.FOCUS_FUTURES)}")
|
|
|
+
|
|
|
+ print(f"生成图表: {'是' if cls.GENERATE_CHARTS else '否'}")
|
|
|
+ print(f"最大分析品种数: {'所有' if cls.MAX_ANALYSIS_FUTURES == -1 else cls.MAX_ANALYSIS_FUTURES}")
|
|
|
+
|
|
|
+# =====================================================================================
|
|
|
+# 第二部分：期货技术形态分析器主类
|
|
|
+# 实现完整的技术形态识别和分析功能
|
|
|
+# =====================================================================================
|
|
|
+
|
|
|
+class FuturePatternAnalyzer:
|
|
|
+ """期货技术形态分析器 - 主要分析类"""
|
|
|
+
|
|
|
+ def __init__(self, start_date=None, end_date=None, config=None):
|
|
|
+ """
|
|
|
+ 初始化分析器
|
|
|
+
|
|
|
+ 参数:
|
|
|
+ - start_date: 分析开始日期，默认使用配置
|
|
|
+ - end_date: 分析结束日期，默认使用配置
|
|
|
+ - config: 自定义配置，默认使用AnalysisConfig
|
|
|
+ """
|
|
|
+ # 使用提供的配置或默认配置
|
|
|
+ if config is None:
|
|
|
+ config = AnalysisConfig.get_config_dict()
|
|
|
+
|
|
|
+ self.start_date = pd.to_datetime(start_date or config['time_range']['start_date'])
|
|
|
+ self.end_date = pd.to_datetime(end_date or config['time_range']['end_date'])
|
|
|
+ self.config = config
|
|
|
+ self.future_codes = {} # 期货代码映射
|
|
|
+
|
|
|
+ # 期货品种与交易所的映射关系
|
|
|
+ self.exchange_map = {
|
|
|
+ # 大连商品交易所 (XDCE)
|
|
|
+ 'A': 'XDCE', 'B': 'XDCE', 'C': 'XDCE', 'CS': 'XDCE', 'FB': 'XDCE',
|
|
|
+ 'I': 'XDCE', 'J': 'XDCE', 'JD': 'XDCE', 'JM': 'XDCE', 'L': 'XDCE',
|
|
|
+ 'M': 'XDCE', 'P': 'XDCE', 'PP': 'XDCE', 'PG': 'XDCE', 'RR': 'XDCE',
|
|
|
+ 'V': 'XDCE', 'Y': 'XDCE', 'EB': 'XDCE', 'EG': 'XDCE', 'LH': 'XDCE',
|
|
|
+
|
|
|
+ # 郑州商品交易所 (XZCE)
|
|
|
+ 'AP': 'XZCE', 'CF': 'XZCE', 'CY': 'XZCE', 'FG': 'XZCE', 'JR': 'XZCE',
|
|
|
+ 'LR': 'XZCE', 'MA': 'XZCE', 'OI': 'XZCE', 'PM': 'XZCE', 'RI': 'XZCE',
|
|
|
+ 'RM': 'XZCE', 'RS': 'XZCE', 'SF': 'XZCE', 'SM': 'XZCE', 'SR': 'XZCE',
|
|
|
+ 'TA': 'XZCE', 'WH': 'XZCE', 'ZC': 'XZCE', 'CJ': 'XZCE',
|
|
|
+ 'ME': 'XZCE', 'PF': 'XZCE', 'PK': 'XZCE', 'RO': 'XZCE', 'SA': 'XZCE',
|
|
|
+ 'TC': 'XZCE', 'UR': 'XZCE', 'WS': 'XZCE', 'WT': 'XZCE',
|
|
|
+
|
|
|
+ # 上海期货交易所 (XSGE)
|
|
|
+ 'AG': 'XSGE', 'AL': 'XSGE', 'AU': 'XSGE', 'BU': 'XSGE', 'CU': 'XSGE',
|
|
|
+ 'FU': 'XSGE', 'HC': 'XSGE', 'NI': 'XSGE', 'PB': 'XSGE', 'RB': 'XSGE',
|
|
|
+ 'RU': 'XSGE', 'SN': 'XSGE', 'SP': 'XSGE', 'SS': 'XSGE', 'WR': 'XSGE',
|
|
|
+ 'ZN': 'XSGE',
|
|
|
+
|
|
|
+ # 上海国际能源交易中心 (XINE)
|
|
|
+ 'BC': 'XINE', 'LU': 'XINE', 'NR': 'XINE', 'SC': 'XINE',
|
|
|
+
|
|
|
+ # 中金所 (CCFX)
|
|
|
+ 'IC': 'CCFX', 'IF': 'CCFX', 'IH': 'CCFX', 'T': 'CCFX', 'TF': 'CCFX', 'TS': 'CCFX',
|
|
|
+
|
|
|
+ # 广期所 (GFEX)
|
|
|
+ 'SI': 'GFEX', 'LC': 'GFEX', 'PS': 'GFEX'
|
|
|
+ }
|
|
|
+
|
|
|
+ self._log(f"初始化分析器 - 分析期间: {self.start_date.strftime('%Y-%m-%d')} 至 {self.end_date.strftime('%Y-%m-%d')}")
|
|
|
+
|
|
|
+ def _log(self, message):
|
|
|
+ """记录日志"""
|
|
|
+ if self.config['analysis_control']['verbose']:
|
|
|
+ print(message)
|
|
|
+
|
|
|
+ def build_future_codes(self):
|
|
|
+ """构建期货品种基础数据"""
|
|
|
+ self._log("\n=== 步骤1: 构建期货品种基础数据 ===")
|
|
|
+
|
|
|
+ # 构建8888主连合约代码
|
|
|
+ for code, exchange in self.exchange_map.items():
|
|
|
+ future_code = f"{code}8888.{exchange}"
|
|
|
+
|
|
|
+ # 应用过滤器
|
|
|
+ if future_code in self.config['future_filter']['excluded']:
|
|
|
+ continue
|
|
|
+
|
|
|
+ # 如果设置了重点关注品种，只添加重点品种
|
|
|
+ if (self.config['future_filter']['focus_only'] and
|
|
|
+ future_code not in self.config['future_filter']['focus_only']):
|
|
|
+ continue
|
|
|
+
|
|
|
+ self.future_codes[code] = future_code
|
|
|
+
|
|
|
+ self._log(f"构建完成，共{len(self.future_codes)}个期货品种")
|
|
|
+ if self.config['analysis_control']['verbose']:
|
|
|
+ self._log("示例品种:")
|
|
|
+ for i, (code, full_code) in enumerate(list(self.future_codes.items())[:5]):
|
|
|
+ self._log(f" {code} -> {full_code}")
|
|
|
+
|
|
|
+ return self.future_codes
|
|
|
+
|
|
|
+ def get_future_price_data(self, future_code, start_date, end_date):
|
|
|
+ """获取期货价格数据"""
|
|
|
+ try:
|
|
|
+ # print(f"获取{future_code}数据: {start_date} 至 {end_date}")
|
|
|
+ data = get_price(
|
|
|
+ future_code,
|
|
|
+ start_date=start_date,
|
|
|
+ end_date=end_date,
|
|
|
+ frequency='daily',
|
|
|
+ fields=['open', 'close', 'high', 'low', 'volume'],
|
|
|
+ skip_paused=False,
|
|
|
+ panel=False
|
|
|
+ )
|
|
|
+ # print(f"获取的最后一条数据: {data.iloc[-1]}")
|
|
|
+ return data
|
|
|
+ except Exception as e:
|
|
|
+ if self.config['analysis_control']['verbose']:
|
|
|
+ print(f" 获取{future_code}数据失败: {str(e)}")
|
|
|
+ return None
|
|
|
+
|
|
|
+ def calculate_drawdown(self, data):
|
|
|
+ """计算最大回撤"""
|
|
|
+ if data is None or len(data) == 0:
|
|
|
+ return None, None, None, None
|
|
|
+
|
|
|
+ # 计算累计最高价
|
|
|
+ data['cum_max'] = data['high'].expanding().max()
|
|
|
+
|
|
|
+ # 计算回撤
|
|
|
+ data['drawdown'] = (data['low'] - data['cum_max']) / data['cum_max']
|
|
|
+
|
|
|
+ # 找到最大回撤
|
|
|
+ max_drawdown = data['drawdown'].min() # 最大回撤是负值，所以用min
|
|
|
+ max_drawdown_idx = data['drawdown'].idxmin()
|
|
|
+
|
|
|
+ # 找到最大回撤对应的高点
|
|
|
+ max_high_before_dd = data.loc[:max_drawdown_idx, 'cum_max'].iloc[-1]
|
|
|
+ high_point_idx = data[data['high'] == max_high_before_dd].index[0]
|
|
|
+
|
|
|
+ return abs(max_drawdown), high_point_idx, max_drawdown_idx, data
|
|
|
+
|
|
|
+ def find_major_decline_futures(self, min_drawdown=None):
|
|
|
+ """筛选大幅下跌的期货品种"""
|
|
|
+ if min_drawdown is None:
|
|
|
+ min_drawdown = self.config['filter_conditions']['min_drawdown']
|
|
|
+
|
|
|
+ self._log(f"\n=== 步骤2: 筛选跌幅>{min_drawdown*100}%的期货品种 ===")
|
|
|
+
|
|
|
+ decline_futures = []
|
|
|
+ total_count = len(self.future_codes)
|
|
|
+ max_futures = self.config['analysis_control']['max_futures']
|
|
|
+ analyzed_count = 0
|
|
|
+
|
|
|
+ for code, future_code in self.future_codes.items():
|
|
|
+ analyzed_count += 1
|
|
|
+
|
|
|
+ # 最大品种数限制
|
|
|
+ if max_futures > 0 and analyzed_count > max_futures:
|
|
|
+ break
|
|
|
+
|
|
|
+ if self.config['analysis_control']['show_progress']:
|
|
|
+ display_total = min(total_count, max_futures) if max_futures > 0 else total_count
|
|
|
+ print(f"分析 {code}({future_code}) [{analyzed_count}/{display_total}]", end=" ")
|
|
|
+
|
|
|
+ # 获取价格数据
|
|
|
+ price_data = self.get_future_price_data(future_code, self.start_date, self.end_date)
|
|
|
+
|
|
|
+ if price_data is None or len(price_data) < self.config['data_quality']['min_data_length']:
|
|
|
+ if self.config['analysis_control']['show_progress']:
|
|
|
+ print("- 数据不足")
|
|
|
+ continue
|
|
|
+
|
|
|
+ # 计算最大回撤
|
|
|
+ max_drawdown, high_date, low_date, data_with_dd = self.calculate_drawdown(price_data)
|
|
|
+
|
|
|
+ if max_drawdown is None or max_drawdown < min_drawdown:
|
|
|
+ if self.config['analysis_control']['show_progress']:
|
|
|
+ print(f"- 最大回撤{max_drawdown:.1%}不符合条件")
|
|
|
+ continue
|
|
|
+
|
|
|
+ # 记录符合条件的品种
|
|
|
+ decline_info = {
|
|
|
+ '品种代码': code,
|
|
|
+ '合约代码': future_code,
|
|
|
+ '最大回撤': max_drawdown,
|
|
|
+ '高点日期': high_date,
|
|
|
+ '低点日期': low_date,
|
|
|
+ '高点价格': data_with_dd.loc[high_date, 'high'],
|
|
|
+ '低点价格': data_with_dd.loc[low_date, 'low'],
|
|
|
+ '下跌天数': (low_date - high_date).days
|
|
|
+ }
|
|
|
+ decline_futures.append(decline_info)
|
|
|
+
|
|
|
+ if self.config['analysis_control']['show_progress']:
|
|
|
+ print(f"✓ 最大回撤{max_drawdown:.1%}, 下跌期间{high_date.strftime('%Y-%m-%d')}至{low_date.strftime('%Y-%m-%d')}")
|
|
|
+
|
|
|
+ self._log(f"\n找到{len(decline_futures)}个符合条件的期货品种")
|
|
|
+ return decline_futures
|
|
|
+
|
|
|
+ def get_simplified_contract_mapping(self, decline_futures):
|
|
|
+ """获取8888加权合约映射 - 简化版本，直接使用加权合约"""
|
|
|
+ self._log("\n=== 步骤3: 创建8888加权合约映射 ===")
|
|
|
+
|
|
|
+ contract_mapping = {}
|
|
|
+
|
|
|
+ for future_info in decline_futures:
|
|
|
+ code = future_info['品种代码']
|
|
|
+ high_date = future_info['高点日期']
|
|
|
+ low_date = future_info['低点日期']
|
|
|
+
|
|
|
+ # 扩展日期范围前后30天
|
|
|
+ extend_days = self.config['analysis_params']['extend_days']
|
|
|
+ start_date = high_date - pd.Timedelta(days=extend_days)
|
|
|
+ end_date = low_date + pd.Timedelta(days=extend_days)
|
|
|
+
|
|
|
+ # 构建8888加权合约代码
|
|
|
+ contract_8888 = self.future_codes.get(code)
|
|
|
+
|
|
|
+ if contract_8888:
|
|
|
+ # 直接使用8888合约，时间范围为分析期间
|
|
|
+ contract_mapping[contract_8888] = [start_date, end_date]
|
|
|
+
|
|
|
+ self._log(f" ✓ {code}: 使用8888加权合约 {contract_8888}")
|
|
|
+ self._log(f" 时间范围: {start_date.strftime('%Y-%m-%d')} 至 {end_date.strftime('%Y-%m-%d')}")
|
|
|
+ else:
|
|
|
+ self._log(f" ⚠️ {code}: 未找到对应的8888合约代码")
|
|
|
+
|
|
|
+ self._log(f"完成8888合约映射，共{len(contract_mapping)}个合约")
|
|
|
+ return contract_mapping
|
|
|
+
|
|
|
+
|
|
|
+ def get_weekly_kline_data(self, contract_mapping):
|
|
|
+ """获取周线K线数据 - 简化版本，直接使用8888加权合约"""
|
|
|
+ self._log("\n=== 步骤4: 获取周线K线数据 ===")
|
|
|
+
|
|
|
+ weekly_data = {}
|
|
|
+
|
|
|
+ for contract_code, time_range in contract_mapping.items():
|
|
|
+ start_date, end_date = time_range
|
|
|
+
|
|
|
+ self._log(f"获取{contract_code}周线数据 ({start_date.strftime('%Y-%m-%d')} 至 {end_date.strftime('%Y-%m-%d')})")
|
|
|
+
|
|
|
+ try:
|
|
|
+ # 计算需要获取的周数，确保充足的数据覆盖
|
|
|
+ time_diff = end_date - start_date
|
|
|
+ estimated_weeks = int(time_diff.days / 7) + 20 # 额外增加20周缓冲
|
|
|
+ estimated_weeks = max(estimated_weeks, 40) # 最少40周
|
|
|
+
|
|
|
+ self._log(f" 计算获取{estimated_weeks}周数据")
|
|
|
+
|
|
|
+ # 直接使用get_bars API获取原生周线数据，避免日线转换的精度损失
|
|
|
+ weekly_bars = get_bars(
|
|
|
+ security=contract_code,
|
|
|
+ count=estimated_weeks,
|
|
|
+ unit='1w', # 直接获取周线数据
|
|
|
+ fields=['date', 'open', 'high', 'low', 'close', 'volume'],
|
|
|
+ include_now=False,
|
|
|
+ end_dt=end_date,
|
|
|
+ fq_ref_date=None,
|
|
|
+ df=True
|
|
|
+ )
|
|
|
+
|
|
|
+ if weekly_bars is not None and len(weekly_bars) > 0:
|
|
|
+ # 确保日期类型一致，避免 'Cannot compare type Timestamp with type date' 错误
|
|
|
+ # get_bars返回的date列可能是date类型，需统一转换为pandas Timestamp
|
|
|
+ weekly_bars['date'] = pd.to_datetime(weekly_bars['date'])
|
|
|
+ start_date = pd.to_datetime(start_date)
|
|
|
+ end_date = pd.to_datetime(end_date)
|
|
|
+
|
|
|
+ # 筛选在指定时间范围内的数据
|
|
|
+ mask = (weekly_bars['date'] >= start_date) & (weekly_bars['date'] <= end_date)
|
|
|
+ filtered_weekly_data = weekly_bars[mask].copy()
|
|
|
+
|
|
|
+ if len(filtered_weekly_data) > 0:
|
|
|
+ # 设置日期为索引
|
|
|
+ filtered_weekly_data.set_index('date', inplace=True)
|
|
|
+
|
|
|
+ # 从合约代码中提取品种代码（用于后续识别）
|
|
|
+ symbol_code = contract_code.replace('8888.', '').split('.')[0]
|
|
|
+
|
|
|
+ weekly_data[symbol_code] = {
|
|
|
+ 'data': filtered_weekly_data,
|
|
|
+ 'contract_code': contract_code,
|
|
|
+ 'time_range': time_range
|
|
|
+ }
|
|
|
+ self._log(f" ✓ {symbol_code}({contract_code}): 获取{len(filtered_weekly_data)}根周K线")
|
|
|
+ else:
|
|
|
+ self._log(f" ⚠️ {contract_code}: 筛选后周线数据不足")
|
|
|
+ else:
|
|
|
+ self._log(f" ⚠️ {contract_code}: 未获取到周线数据")
|
|
|
+
|
|
|
+ except Exception as e:
|
|
|
+ self._log(f" ❌ 获取{contract_code}周线数据失败: {str(e)}")
|
|
|
+ continue
|
|
|
+
|
|
|
+ self._log(f"完成周线数据获取，共{len(weekly_data)}个品种")
|
|
|
+ return weekly_data
|
|
|
+
|
|
|
+ def get_weekly_data_by_contract_log(self, contract_switch_log, extend_weeks=10):
|
|
|
+ """
|
|
|
+ 基于合约切换日志获取每个合约的周线数据
|
|
|
+
|
|
|
+ 参数:
|
|
|
+ - contract_switch_log: 合约切换日志，格式为列表，每个元素包含：
|
|
|
+ {'contract': '合约代码', 'start_date': '开始日期', 'end_date': '结束日期'}
|
|
|
+ - extend_weeks: 延长周数，默认10周
|
|
|
+
|
|
|
+ 返回:
|
|
|
+ - dict: 每个合约的周线数据，键为合约代码，值为数据和时间范围信息
|
|
|
+ """
|
|
|
+ self._log(f"\n=== 基于合约切换日志获取周线数据 ===")
|
|
|
+ self._log(f"延长周数: {extend_weeks}周")
|
|
|
+
|
|
|
+ weekly_data_by_contract = {}
|
|
|
+
|
|
|
+ for contract_info in contract_switch_log:
|
|
|
+ contract_code = contract_info.get('contract')
|
|
|
+ start_date = contract_info.get('start_date')
|
|
|
+ end_date = contract_info.get('end_date')
|
|
|
+
|
|
|
+ if not all([contract_code, start_date, end_date]):
|
|
|
+ self._log(f"⚠️ 合约信息不完整，跳过: {contract_info}")
|
|
|
+ continue
|
|
|
+
|
|
|
+ # 转换日期格式
|
|
|
+ if isinstance(start_date, str):
|
|
|
+ start_date = pd.to_datetime(start_date)
|
|
|
+ if isinstance(end_date, str):
|
|
|
+ end_date = pd.to_datetime(end_date)
|
|
|
+
|
|
|
+ # 计算延长后的时间范围
|
|
|
+ extended_start_date = start_date - pd.Timedelta(weeks=extend_weeks)
|
|
|
+ extended_end_date = end_date + pd.Timedelta(weeks=extend_weeks)
|
|
|
+
|
|
|
+ self._log(f"处理合约: {contract_code}")
|
|
|
+ self._log(f" 原始时间范围: {start_date.strftime('%Y-%m-%d')} 至 {end_date.strftime('%Y-%m-%d')}")
|
|
|
+ self._log(f" 延长时间范围: {extended_start_date.strftime('%Y-%m-%d')} 至 {extended_end_date.strftime('%Y-%m-%d')}")
|
|
|
+
|
|
|
+ try:
|
|
|
+ # 计算需要获取的周数，采用更稳健的方式
|
|
|
+ time_diff = extended_end_date - extended_start_date
|
|
|
+ estimated_weeks = int(time_diff.days / 7) + 20 # 额外增加20周作为充足缓冲
|
|
|
+
|
|
|
+ # 确保最少获取60周数据（约1.2年）以提供充分的历史背景
|
|
|
+ estimated_weeks = max(estimated_weeks, 60)
|
|
|
+
|
|
|
+ self._log(f" 计算获取{estimated_weeks}周数据 (时间跨度{time_diff.days}天)")
|
|
|
+
|
|
|
+ # 直接使用get_bars API获取原生周线数据，避免日线转换的精度损失
|
|
|
+ weekly_bars = get_bars(
|
|
|
+ security=contract_code,
|
|
|
+ count=estimated_weeks, # 扩展的周数以确保完整覆盖
|
|
|
+ unit='1w', # 直接获取周线数据，避免转换精度损失
|
|
|
+ fields=['date', 'open', 'high', 'low', 'close', 'volume'],
|
|
|
+ include_now=False, # 不包含当前未完成的周期
|
|
|
+ end_dt=extended_end_date,
|
|
|
+ fq_ref_date=None, # 不复权，保持原始价格
|
|
|
+ df=True # 返回DataFrame格式便于处理
|
|
|
+ )
|
|
|
+
|
|
|
+ if weekly_bars is not None and len(weekly_bars) > 0:
|
|
|
+ # 确保日期类型一致，避免 'Cannot compare type Timestamp with type date' 错误
|
|
|
+ # get_bars返回的date列可能是date类型，需统一转换为pandas Timestamp
|
|
|
+ weekly_bars['date'] = pd.to_datetime(weekly_bars['date'])
|
|
|
+ extended_start_date = pd.to_datetime(extended_start_date)
|
|
|
+ extended_end_date = pd.to_datetime(extended_end_date)
|
|
|
+
|
|
|
+ # 筛选在延长时间范围内的数据
|
|
|
+ mask = (weekly_bars['date'] >= extended_start_date) & (weekly_bars['date'] <= extended_end_date)
|
|
|
+ filtered_data = weekly_bars[mask].copy()
|
|
|
+
|
|
|
+ if len(filtered_data) > 0:
|
|
|
+ # 设置日期为索引
|
|
|
+ filtered_data.set_index('date', inplace=True)
|
|
|
+
|
|
|
+ # 存储数据
|
|
|
+ weekly_data_by_contract[contract_code] = {
|
|
|
+ 'data': filtered_data,
|
|
|
+ 'original_start': start_date,
|
|
|
+ 'original_end': end_date,
|
|
|
+ 'extended_start': extended_start_date,
|
|
|
+ 'extended_end': extended_end_date,
|
|
|
+ 'total_weeks': len(filtered_data)
|
|
|
+ }
|
|
|
+
|
|
|
+ self._log(f" ✓ 获取到{len(filtered_data)}根周K线数据")
|
|
|
+ self._log(f" 数据时间范围: {filtered_data.index[0].strftime('%Y-%m-%d')} 至 {filtered_data.index[-1].strftime('%Y-%m-%d')}")
|
|
|
+
|
|
|
+ # 添加数据质量检查
|
|
|
+ original_start_ts = pd.to_datetime(start_date)
|
|
|
+ original_end_ts = pd.to_datetime(end_date)
|
|
|
+ original_weeks = int((original_end_ts - original_start_ts).days / 7) + 1
|
|
|
+ expected_total_weeks = original_weeks + 2 * extend_weeks
|
|
|
+ if len(filtered_data) < expected_total_weeks * 0.7: # 如果数据不足预期的70%
|
|
|
+ self._log(f" ⚠️ 数据可能不完整，预期约{expected_total_weeks}周，实际{len(filtered_data)}周")
|
|
|
+ else:
|
|
|
+ self._log(f" ⚠️ 筛选后无有效数据")
|
|
|
+
|
|
|
+ else:
|
|
|
+ self._log(f" ⚠️ 未获取到周线数据")
|
|
|
+
|
|
|
+ except Exception as e:
|
|
|
+ self._log(f" ❌ 获取{contract_code}周线数据失败: {str(e)}")
|
|
|
+ continue
|
|
|
+
|
|
|
+ self._log(f"\n成功获取{len(weekly_data_by_contract)}个合约的周线数据")
|
|
|
+ return weekly_data_by_contract
|
|
|
+
|
|
|
+ def identify_target_pattern(self, weekly_data, decline_futures):
|
|
|
+ """
|
|
|
+ 识别目标K线形态
|
|
|
+
|
|
|
+ 当未识别到任何目标形态时，会自动显示完整的烛台图（K线图）作为数据可视化备选方案。
|
|
|
+ 图表包含完整的OHLC数据，直接显示不保存到本地存储。
|
|
|
+ """
|
|
|
+ self._log("\n=== 步骤5: 识别目标K线形态 ===")
|
|
|
+
|
|
|
+ pattern_results = []
|
|
|
+ kline_conditions = self.config['filter_conditions']['kline_conditions']
|
|
|
+
|
|
|
+ # 创建品种代码到下跌信息的映射
|
|
|
+ decline_info_map = {future['品种代码']: future for future in decline_futures}
|
|
|
+
|
|
|
+ for code, data_info in weekly_data.items():
|
|
|
+ data = data_info['data']
|
|
|
+ contract_code = data_info['contract_code'] # 8888加权合约代码
|
|
|
+
|
|
|
+ # 从decline_futures中获取下跌时间信息
|
|
|
+ if code not in decline_info_map:
|
|
|
+ continue
|
|
|
+
|
|
|
+ decline_info = decline_info_map[code]
|
|
|
+ decline_start = decline_info['高点日期']
|
|
|
+ decline_end = decline_info['低点日期']
|
|
|
+ extend_days = self.config['analysis_params']['extend_days']
|
|
|
+
|
|
|
+ self._log(f"分析{code}的K线形态 (使用{contract_code})")
|
|
|
+
|
|
|
+ for idx, (date, row) in enumerate(data.iterrows()):
|
|
|
+ open_price = row['open']
|
|
|
+ close_price = row['close']
|
|
|
+ high_price = row['high']
|
|
|
+ low_price = row['low']
|
|
|
+ volume = row['volume']
|
|
|
+
|
|
|
+ # 检查是否在下跌期间或附近
|
|
|
+ date_in_range = (date >= decline_start - pd.Timedelta(days=extend_days) and
|
|
|
+ date <= decline_end + pd.Timedelta(days=extend_days))
|
|
|
+
|
|
|
+ if not date_in_range:
|
|
|
+ continue
|
|
|
+
|
|
|
+ # 使用统一的核心验证逻辑
|
|
|
+ validation_result = self._core_pattern_validation(date, row, kline_conditions)
|
|
|
+
|
|
|
+ if validation_result['validation_passed']:
|
|
|
+ pattern_info = {
|
|
|
+ '品种代码': code,
|
|
|
+ '日期': date,
|
|
|
+ '开盘价': validation_result['ohlc_data']['open'],
|
|
|
+ '收盘价': validation_result['ohlc_data']['close'],
|
|
|
+ '最高价': validation_result['ohlc_data']['high'],
|
|
|
+ '最低价': validation_result['ohlc_data']['low'],
|
|
|
+ '成交量': validation_result['ohlc_data']['volume'],
|
|
|
+ '开收差异率': validation_result['indicators']['price_diff_pct']['actual_value'],
|
|
|
+ '最高涨幅率': validation_result['indicators']['high_gain_pct']['actual_value'],
|
|
|
+ '下跌开始日期': decline_start,
|
|
|
+ '下跌结束日期': decline_end,
|
|
|
+ '数据索引': idx,
|
|
|
+ '使用合约': contract_code # 8888加权合约
|
|
|
+ }
|
|
|
+ pattern_results.append(pattern_info)
|
|
|
+
|
|
|
+ price_diff_pct = validation_result['indicators']['price_diff_pct']['actual_value'] / 100
|
|
|
+ high_gain_pct = validation_result['indicators']['high_gain_pct']['actual_value'] / 100
|
|
|
+ self._log(f" ✓ 找到形态: {date.strftime('%Y-%m-%d')} 开收差{price_diff_pct:.1%} 最高涨{high_gain_pct:.1%} 合约{contract_code}")
|
|
|
+
|
|
|
+ self._log(f"\n共识别出{len(pattern_results)}个目标K线形态")
|
|
|
+
|
|
|
+ # 如果未识别到任何模式，显示完整的烛台图
|
|
|
+ if len(pattern_results) == 0:
|
|
|
+ self._log("\n⚠️ 未识别到任何目标形态，显示完整数据集的K线图")
|
|
|
+ self._display_full_candlestick_charts(weekly_data)
|
|
|
+
|
|
|
+ return pattern_results
|
|
|
+
|
|
|
+ def _display_full_candlestick_charts(self, weekly_data):
|
|
|
+ """显示完整的烛台图（K线图）"""
|
|
|
+ self._log("=== 显示完整数据集的K线图 ===")
|
|
|
+
|
|
|
+ for code, data_info in weekly_data.items():
|
|
|
+ data = data_info['data']
|
|
|
+ contract_code = data_info['contract_code']
|
|
|
+
|
|
|
+ if data is None or len(data) == 0:
|
|
|
+ self._log(f"⚠️ {code}: 无可用数据")
|
|
|
+ continue
|
|
|
+
|
|
|
+ self._log(f"绘制{code}({contract_code})的完整K线图，共{len(data)}根K线")
|
|
|
+
|
|
|
+ try:
|
|
|
+ # 创建图表
|
|
|
+ fig, ax = plt.subplots(figsize=(16, 10))
|
|
|
+
|
|
|
+ # 准备数据用于绘制烛台图
|
|
|
+ dates = data.index
|
|
|
+ opens = data['open']
|
|
|
+ highs = data['high']
|
|
|
+ lows = data['low']
|
|
|
+ closes = data['close']
|
|
|
+ volumes = data['volume']
|
|
|
+
|
|
|
+ # 绘制K线图
|
|
|
+ for i in range(len(data)):
|
|
|
+ date = dates[i]
|
|
|
+ open_price = opens.iloc[i]
|
|
|
+ high_price = highs.iloc[i]
|
|
|
+ low_price = lows.iloc[i]
|
|
|
+ close_price = closes.iloc[i]
|
|
|
+
|
|
|
+ # K线颜色：红涨绿跌
|
|
|
+ color = 'red' if close_price > open_price else 'green'
|
|
|
+ edge_color = 'darkred' if close_price > open_price else 'darkgreen'
|
|
|
+
|
|
|
+ # 绘制影线（最高价到最低价的竖线）
|
|
|
+ ax.plot([i, i], [low_price, high_price], color='black', linewidth=1)
|
|
|
+
|
|
|
+ # 绘制实体（开盘价到收盘价的矩形）
|
|
|
+ body_height = abs(close_price - open_price)
|
|
|
+ bottom = min(open_price, close_price)
|
|
|
+
|
|
|
+ # 使用矩形绘制K线实体
|
|
|
+ rect = patches.Rectangle((i-0.4, bottom), 0.8, body_height,
|
|
|
+ linewidth=1, edgecolor=edge_color,
|
|
|
+ facecolor=color, alpha=0.8)
|
|
|
+ ax.add_patch(rect)
|
|
|
+
|
|
|
+ # 设置图表标题和标签
|
|
|
+ ax.set_title(f'{code} ({contract_code}) 完整周线K线图\n'
|
|
|
+ f'数据期间: {dates[0].strftime("%Y-%m-%d")} 至 {dates[-1].strftime("%Y-%m-%d")} '
|
|
|
+ f'(共{len(data)}根K线)',
|
|
|
+ fontsize=14, fontweight='bold', pad=20)
|
|
|
+
|
|
|
+ ax.set_xlabel('时间', fontsize=12)
|
|
|
+ ax.set_ylabel('价格', fontsize=12)
|
|
|
+ ax.grid(True, alpha=0.3)
|
|
|
+
|
|
|
+ # 设置x轴标签
|
|
|
+ # 选择合适的标签间隔，避免过于密集
|
|
|
+ step = max(1, len(data) // 10) # 显示约10个时间标签
|
|
|
+ tick_positions = range(0, len(data), step)
|
|
|
+ tick_labels = [dates[i].strftime('%Y-%m-%d') for i in tick_positions]
|
|
|
+
|
|
|
+ ax.set_xticks(tick_positions)
|
|
|
+ ax.set_xticklabels(tick_labels, rotation=45, ha='right')
|
|
|
+
|
|
|
+ # 添加统计信息
|
|
|
+ max_price = highs.max()
|
|
|
+ min_price = lows.min()
|
|
|
+ latest_close = closes.iloc[-1]
|
|
|
+ total_change = (latest_close - closes.iloc[0]) / closes.iloc[0] * 100
|
|
|
+
|
|
|
+ # 在图表上添加统计文本
|
|
|
+ stats_text = (f'最高价: {max_price:.2f}\n'
|
|
|
+ f'最低价: {min_price:.2f}\n'
|
|
|
+ f'最新收盘: {latest_close:.2f}\n'
|
|
|
+ f'总涨跌幅: {total_change:+.2f}%')
|
|
|
+
|
|
|
+ ax.text(0.02, 0.98, stats_text, transform=ax.transAxes,
|
|
|
+ verticalalignment='top', bbox=dict(boxstyle='round',
|
|
|
+ facecolor='wheat', alpha=0.8), fontsize=10)
|
|
|
+
|
|
|
+ # 调整布局并显示
|
|
|
+ plt.tight_layout()
|
|
|
+ plt.show()
|
|
|
+
|
|
|
+ self._log(f"✓ 已显示{code}的完整K线图")
|
|
|
+
|
|
|
+ except Exception as e:
|
|
|
+ self._log(f"❌ 绘制{code}K线图时出错: {str(e)}")
|
|
|
+ continue
|
|
|
+
|
|
|
+ self._log("完成所有品种的K线图显示")
|
|
|
+
|
|
|
+ def _core_pattern_validation(self, target_date, kline_data_row, kline_conditions):
|
|
|
+ """
|
|
|
+ 统一的核心模式验证逻辑
|
|
|
+
|
|
|
+ 此函数包含两种模式共同的验证逻辑，消除代码重复。
|
|
|
+ 无论是自动模式遍历所有K线，还是验证模式检查特定日期，
|
|
|
+ 都使用相同的验证标准和计算方法。
|
|
|
+
|
|
|
+ 参数:
|
|
|
+ - target_date: 目标日期 (pandas Timestamp)
|
|
|
+ - kline_data_row: 单根K线数据 (pandas Series，包含open、high、low、close、volume)
|
|
|
+ - kline_conditions: 验证条件配置字典
|
|
|
+
|
|
|
+ 返回:
|
|
|
+ - dict: 包含验证结果和详细诊断信息的字典
|
|
|
+ """
|
|
|
+
|
|
|
+ # 提取OHLC数据
|
|
|
+ open_price = kline_data_row['open']
|
|
|
+ close_price = kline_data_row['close']
|
|
|
+ high_price = kline_data_row['high']
|
|
|
+ low_price = kline_data_row['low']
|
|
|
+ volume = kline_data_row['volume']
|
|
|
+
|
|
|
+ # 计算模式指标
|
|
|
+ price_diff_pct = abs(close_price - open_price) / open_price
|
|
|
+ high_gain_pct = (high_price - open_price) / open_price
|
|
|
+
|
|
|
+ # 验证条件A: 开盘价与收盘价差异
|
|
|
+ condition_a = price_diff_pct <= kline_conditions['max_open_close_diff']
|
|
|
+ condition_a_threshold = kline_conditions['max_open_close_diff']
|
|
|
+
|
|
|
+ # 验证条件B: 最高价相对开盘价涨幅
|
|
|
+ condition_b = high_gain_pct >= kline_conditions['min_high_open_gain']
|
|
|
+ condition_b_threshold = kline_conditions['min_high_open_gain']
|
|
|
+
|
|
|
+ # 整体验证结果
|
|
|
+ validation_passed = condition_a and condition_b
|
|
|
+
|
|
|
+ # 构建详细的验证结果
|
|
|
+ result = {
|
|
|
+ 'date': target_date,
|
|
|
+ 'validation_passed': validation_passed,
|
|
|
+ 'ohlc_data': {
|
|
|
+ 'open': round(open_price, 2),
|
|
|
+ 'high': round(high_price, 2),
|
|
|
+ 'low': round(low_price, 2),
|
|
|
+ 'close': round(close_price, 2),
|
|
|
+ 'volume': volume
|
|
|
+ },
|
|
|
+ 'indicators': {
|
|
|
+ 'price_diff_pct': {
|
|
|
+ 'actual_value': round(price_diff_pct * 100, 2),
|
|
|
+ 'threshold': round(condition_a_threshold * 100, 2),
|
|
|
+ 'unit': '%',
|
|
|
+ 'condition': '≤',
|
|
|
+ 'passed': condition_a,
|
|
|
+ 'description': '开盘价与收盘价差异率'
|
|
|
+ },
|
|
|
+ 'high_gain_pct': {
|
|
|
+ 'actual_value': round(high_gain_pct * 100, 2),
|
|
|
+ 'threshold': round(condition_b_threshold * 100, 2),
|
|
|
+ 'unit': '%',
|
|
|
+ 'condition': '≥',
|
|
|
+ 'passed': condition_b,
|
|
|
+ 'description': '最高价相对开盘价涨幅'
|
|
|
+ }
|
|
|
+ }
|
|
|
+ }
|
|
|
+
|
|
|
+ # 生成失败原因
|
|
|
+ if not validation_passed:
|
|
|
+ failed_conditions = []
|
|
|
+ if not condition_a:
|
|
|
+ failed_conditions.append(f"开收差异率{price_diff_pct*100:.2f}% > 阈值{condition_a_threshold*100:.1f}%")
|
|
|
+ if not condition_b:
|
|
|
+ failed_conditions.append(f"最高涨幅{high_gain_pct*100:.2f}% < 阈值{condition_b_threshold*100:.1f}%")
|
|
|
+
|
|
|
+ result['failure_reason'] = "; ".join(failed_conditions)
|
|
|
+
|
|
|
+ return result
|
|
|
+
|
|
|
+ def validate_manual_patterns(self, manual_config, weekly_data):
|
|
|
+ """
|
|
|
+ 手动验证模式功能
|
|
|
+
|
|
|
+ 参数:
|
|
|
+ - manual_config: 手动配置字典，格式：
|
|
|
+ {'B':['2022-07-22','2024-08-23'], 'A':['2023-01-06','2023-06-09']}
|
|
|
+ - weekly_data: 周线数据字典
|
|
|
+
|
|
|
+ 返回:
|
|
|
+ - dict: 验证结果，包含成功和失败的详细信息
|
|
|
+ """
|
|
|
+ self._log("\n=== 手动模式验证 ===")
|
|
|
+ self._log(f"配置的模式类型数: {len(manual_config)}")
|
|
|
+
|
|
|
+ validation_results = {
|
|
|
+ 'successful_validations': [],
|
|
|
+ 'failed_validations': [],
|
|
|
+ 'summary': {}
|
|
|
+ }
|
|
|
+
|
|
|
+ kline_conditions = self.config['filter_conditions']['kline_conditions']
|
|
|
+
|
|
|
+ for pattern_type, date_list in manual_config.items():
|
|
|
+ self._log(f"\n处理模式类型 '{pattern_type}': {len(date_list)} 个日期")
|
|
|
+
|
|
|
+ type_results = {
|
|
|
+ 'pattern_type': pattern_type,
|
|
|
+ 'total_dates': len(date_list),
|
|
|
+ 'successful_count': 0,
|
|
|
+ 'failed_count': 0,
|
|
|
+ 'validations': []
|
|
|
+ }
|
|
|
+
|
|
|
+ for date_str in date_list:
|
|
|
+ try:
|
|
|
+ # 转换日期格式
|
|
|
+ target_date = pd.to_datetime(date_str)
|
|
|
+ self._log(f" 验证日期: {target_date.strftime('%Y-%m-%d')}")
|
|
|
+
|
|
|
+ # 验证这个日期
|
|
|
+ validation_result = self._validate_single_date(
|
|
|
+ target_date, pattern_type, weekly_data, kline_conditions
|
|
|
+ )
|
|
|
+
|
|
|
+ type_results['validations'].append(validation_result)
|
|
|
+
|
|
|
+ if validation_result['validation_passed']:
|
|
|
+ type_results['successful_count'] += 1
|
|
|
+ validation_results['successful_validations'].append(validation_result)
|
|
|
+ self._log(f" ✓ 验证通过")
|
|
|
+ else:
|
|
|
+ type_results['failed_count'] += 1
|
|
|
+ validation_results['failed_validations'].append(validation_result)
|
|
|
+ self._log(f" ❌ 验证失败: {validation_result['failure_reason']}")
|
|
|
+
|
|
|
+ except Exception as e:
|
|
|
+ error_result = {
|
|
|
+ 'pattern_type': pattern_type,
|
|
|
+ 'target_date': date_str,
|
|
|
+ 'validation_passed': False,
|
|
|
+ 'failure_reason': f'日期处理错误: {str(e)}',
|
|
|
+ 'error': True
|
|
|
+ }
|
|
|
+ type_results['validations'].append(error_result)
|
|
|
+ type_results['failed_count'] += 1
|
|
|
+ validation_results['failed_validations'].append(error_result)
|
|
|
+ self._log(f" ❌ 日期处理错误: {str(e)}")
|
|
|
+
|
|
|
+ validation_results['summary'][pattern_type] = type_results
|
|
|
+
|
|
|
+ # 打印总体统计
|
|
|
+ total_validations = sum(result['total_dates'] for result in validation_results['summary'].values())
|
|
|
+ total_successful = len(validation_results['successful_validations'])
|
|
|
+ total_failed = len(validation_results['failed_validations'])
|
|
|
+
|
|
|
+ self._log(f"\n=== 验证结果汇总 ===")
|
|
|
+ self._log(f"总验证数: {total_validations}")
|
|
|
+ self._log(f"成功验证: {total_successful}")
|
|
|
+ self._log(f"失败验证: {total_failed}")
|
|
|
+ self._log(f"成功率: {(total_successful/total_validations*100) if total_validations > 0 else 0:.1f}%")
|
|
|
+
|
|
|
+ # 显示失败详情
|
|
|
+ if validation_results['failed_validations']:
|
|
|
+ self._display_validation_failures(validation_results['failed_validations'])
|
|
|
+
|
|
|
+ return validation_results
|
|
|
+
|
|
|
+ def _validate_single_date(self, target_date, pattern_type, weekly_data, kline_conditions):
|
|
|
+ """验证单个日期的模式 - 使用统一的核心验证逻辑"""
|
|
|
+
|
|
|
+ validation_result = {
|
|
|
+ 'pattern_type': pattern_type,
|
|
|
+ 'target_date': target_date.strftime('%Y-%m-%d'),
|
|
|
+ 'validation_passed': False,
|
|
|
+ 'failure_reason': '',
|
|
|
+ 'diagnostics': {}
|
|
|
+ }
|
|
|
+
|
|
|
+ # 查找包含目标日期的数据
|
|
|
+ found_data = None
|
|
|
+
|
|
|
+ for code, data_info in weekly_data.items():
|
|
|
+ data = data_info['data']
|
|
|
+ contract_code = data_info['contract_code']
|
|
|
+
|
|
|
+ # 检查目标日期是否在数据范围内
|
|
|
+ if target_date >= data.index[0] and target_date <= data.index[-1]:
|
|
|
+ # 找到最接近的周线数据
|
|
|
+ closest_idx = data.index.get_indexer([target_date], method='nearest')[0]
|
|
|
+ if closest_idx >= 0 and closest_idx < len(data):
|
|
|
+ found_data = data.iloc[closest_idx]
|
|
|
+ validation_result['contract_code'] = contract_code
|
|
|
+ validation_result['closest_date'] = data.index[closest_idx].strftime('%Y-%m-%d')
|
|
|
+ break
|
|
|
+
|
|
|
+ if found_data is None:
|
|
|
+ validation_result['failure_reason'] = '未找到匹配的数据'
|
|
|
+ return validation_result
|
|
|
+
|
|
|
+ # 使用统一的核心验证逻辑
|
|
|
+ core_result = self._core_pattern_validation(target_date, found_data, kline_conditions)
|
|
|
+
|
|
|
+ # 整合核心结果到验证结果
|
|
|
+ validation_result['validation_passed'] = core_result['validation_passed']
|
|
|
+ validation_result['diagnostics']['ohlc_data'] = core_result['ohlc_data']
|
|
|
+ validation_result['diagnostics']['indicators'] = core_result['indicators']
|
|
|
+
|
|
|
+ if 'failure_reason' in core_result:
|
|
|
+ validation_result['failure_reason'] = core_result['failure_reason']
|
|
|
+
|
|
|
+ return validation_result
|
|
|
+
|
|
|
+ def _display_validation_failures(self, failed_validations):
|
|
|
+ """显示详细的验证失败信息"""
|
|
|
+ self._log("\n=== 详细失败诊断 ===")
|
|
|
+
|
|
|
+ for i, failure in enumerate(failed_validations, 1):
|
|
|
+ if failure.get('error'):
|
|
|
+ continue # 跳过错误情况
|
|
|
+
|
|
|
+ self._log(f"\n失败 #{i}: 模式类型 '{failure['pattern_type']}' - 日期 {failure['target_date']}")
|
|
|
+
|
|
|
+ if 'contract_code' in failure:
|
|
|
+ self._log(f" 合约: {failure['contract_code']}")
|
|
|
+ self._log(f" 最近日期: {failure['closest_date']}")
|
|
|
+
|
|
|
+ # 显示OHLC数据
|
|
|
+ if 'ohlc_data' in failure['diagnostics']:
|
|
|
+ ohlc = failure['diagnostics']['ohlc_data']
|
|
|
+ self._log(f" OHLC数据:")
|
|
|
+ self._log(f" 开盘: {ohlc['open']:>8}")
|
|
|
+ self._log(f" 最高: {ohlc['high']:>8}")
|
|
|
+ self._log(f" 最低: {ohlc['low']:>8}")
|
|
|
+ self._log(f" 收盘: {ohlc['close']:>8}")
|
|
|
+ self._log(f" 成交量: {ohlc['volume']}")
|
|
|
+
|
|
|
+ # 显示指标验证详情
|
|
|
+ if 'indicators' in failure['diagnostics']:
|
|
|
+ self._log(f" 指标验证:")
|
|
|
+ for indicator_name, indicator_data in failure['diagnostics']['indicators'].items():
|
|
|
+ status = "✓" if indicator_data['passed'] else "❌"
|
|
|
+ self._log(f" {status} {indicator_data['description']}:")
|
|
|
+ self._log(f" 实际值: {indicator_data['actual_value']}{indicator_data['unit']}")
|
|
|
+ self._log(f" 阈值: {indicator_data['condition']} {indicator_data['threshold']}{indicator_data['unit']}")
|
|
|
+ self._log(f" 结果: {'通过' if indicator_data['passed'] else '失败'}")
|
|
|
+
|
|
|
+ self._log(f" 失败原因: {failure['failure_reason']}")
|
|
|
+
|
|
|
+ def analyze_future_performance(self, pattern_results, weekly_data):
|
|
|
+ """分析后续走势 - 逐周详细分析"""
|
|
|
+ self._log("\n=== 步骤6: 分析后续10周逐周表现 ===")
|
|
|
+
|
|
|
+ performance_results = []
|
|
|
+ follow_up_weeks = self.config['analysis_params']['follow_up_weeks']
|
|
|
+
|
|
|
+ for pattern in pattern_results:
|
|
|
+ code = pattern['品种代码']
|
|
|
+ pattern_date = pattern['日期']
|
|
|
+ pattern_idx = pattern['数据索引']
|
|
|
+ pattern_close_price = pattern['收盘价']
|
|
|
+
|
|
|
+ if code not in weekly_data:
|
|
|
+ continue
|
|
|
+
|
|
|
+ data = weekly_data[code]['data']
|
|
|
+
|
|
|
+ # 基础信息
|
|
|
+ future_performance = {
|
|
|
+ '品种代码': code,
|
|
|
+ '形态日期': pattern_date,
|
|
|
+ '形态收盘价': pattern_close_price
|
|
|
+ }
|
|
|
+
|
|
|
+ # 详细的逐周分析
|
|
|
+ print(f"\n{code} ({pattern_date.strftime('%Y-%m-%d')}) 后续10周详细分析:")
|
|
|
+ print("-" * 80)
|
|
|
+ print(f"{'周次':<4} {'日期':<12} {'开盘':<8} {'收盘':<8} {'最高':<8} {'最低':<8} "
|
|
|
+ f"{'周涨跌幅':<10} {'最大跌幅':<10}")
|
|
|
+ print("-" * 80)
|
|
|
+
|
|
|
+ # 逐周分析
|
|
|
+ for weeks in range(1, follow_up_weeks + 1):
|
|
|
+ future_idx = pattern_idx + weeks
|
|
|
+
|
|
|
+ if future_idx < len(data):
|
|
|
+ future_row = data.iloc[future_idx]
|
|
|
+ week_date = data.index[future_idx]
|
|
|
+
|
|
|
+ week_open = future_row['open']
|
|
|
+ week_close = future_row['close']
|
|
|
+ week_high = future_row['high']
|
|
|
+ week_low = future_row['low']
|
|
|
+
|
|
|
+ # a) 每周的涨跌幅：收盘价相对于开盘价的涨跌幅度
|
|
|
+ weekly_change_pct = (week_close - week_open) / week_open * 100
|
|
|
+
|
|
|
+ # b) 每周的最大跌幅：开盘价到当周最低价的最大跌幅
|
|
|
+ max_decline_pct = (week_low - week_open) / week_open * 100
|
|
|
+
|
|
|
+ # 存储详细数据
|
|
|
+ future_performance[f'第{weeks}周_日期'] = week_date.strftime('%Y-%m-%d')
|
|
|
+ future_performance[f'第{weeks}周_开盘价'] = round(week_open, 2)
|
|
|
+ future_performance[f'第{weeks}周_收盘价'] = round(week_close, 2)
|
|
|
+ future_performance[f'第{weeks}周_最高价'] = round(week_high, 2)
|
|
|
+ future_performance[f'第{weeks}周_最低价'] = round(week_low, 2)
|
|
|
+ future_performance[f'第{weeks}周_涨跌幅'] = round(weekly_change_pct, 2)
|
|
|
+ future_performance[f'第{weeks}周_最大跌幅'] = round(max_decline_pct, 2)
|
|
|
+
|
|
|
+ # 打印逐周详细信息
|
|
|
+ print(f"{weeks:>2} {week_date.strftime('%Y-%m-%d'):<12} "
|
|
|
+ f"{week_open:<8.2f} {week_close:<8.2f} {week_high:<8.2f} {week_low:<8.2f} "
|
|
|
+ f"{weekly_change_pct:>+8.2f}% {max_decline_pct:>+8.2f}%")
|
|
|
+
|
|
|
+ else:
|
|
|
+ # 数据不足的情况
|
|
|
+ for field in ['日期', '开盘价', '收盘价', '最高价', '最低价', '涨跌幅', '最大跌幅']:
|
|
|
+ future_performance[f'第{weeks}周_{field}'] = None
|
|
|
+ print(f"{weeks:>2} {'无数据':<12} {'--':<8} {'--':<8} {'--':<8} {'--':<8} "
|
|
|
+ f"{'--':<10} {'--':<10}")
|
|
|
+
|
|
|
+ print("-" * 80)
|
|
|
+
|
|
|
+ # 计算统计指标
|
|
|
+ weekly_changes = [future_performance[f'第{w}周_涨跌幅']
|
|
|
+ for w in range(1, follow_up_weeks + 1)
|
|
|
+ if future_performance[f'第{w}周_涨跌幅'] is not None]
|
|
|
+
|
|
|
+ max_declines = [future_performance[f'第{w}周_最大跌幅']
|
|
|
+ for w in range(1, follow_up_weeks + 1)
|
|
|
+ if future_performance[f'第{w}周_最大跌幅'] is not None]
|
|
|
+
|
|
|
+ if weekly_changes:
|
|
|
+ avg_weekly_change = sum(weekly_changes) / len(weekly_changes)
|
|
|
+ max_weekly_gain = max(weekly_changes)
|
|
|
+ max_weekly_loss = min(weekly_changes)
|
|
|
+
|
|
|
+ future_performance['统计_平均周涨跌幅'] = round(avg_weekly_change, 2)
|
|
|
+ future_performance['统计_最大周涨幅'] = round(max_weekly_gain, 2)
|
|
|
+ future_performance['统计_最大周跌幅'] = round(max_weekly_loss, 2)
|
|
|
+
|
|
|
+ print(f"统计汇总:")
|
|
|
+ print(f" 平均周涨跌幅: {avg_weekly_change:+.2f}%")
|
|
|
+ print(f" 最大周涨幅: {max_weekly_gain:+.2f}%")
|
|
|
+ print(f" 最大周跌幅: {max_weekly_loss:+.2f}%")
|
|
|
+
|
|
|
+ if max_declines:
|
|
|
+ avg_max_decline = sum(max_declines) / len(max_declines)
|
|
|
+ worst_decline = min(max_declines)
|
|
|
+
|
|
|
+ future_performance['统计_平均最大跌幅'] = round(avg_max_decline, 2)
|
|
|
+ future_performance['统计_最差跌幅'] = round(worst_decline, 2)
|
|
|
+
|
|
|
+ print(f" 平均最大跌幅: {avg_max_decline:+.2f}%")
|
|
|
+ print(f" 最差单周跌幅: {worst_decline:+.2f}%")
|
|
|
+
|
|
|
+ performance_results.append(future_performance)
|
|
|
+ print("=" * 80)
|
|
|
+
|
|
|
+ return performance_results
|
|
|
+
|
|
|
+ def create_visualizations(self, pattern_results, weekly_data):
|
|
|
+ """创建K线图可视化"""
|
|
|
+ if not self.config['output_settings']['generate_charts']:
|
|
|
+ return
|
|
|
+
|
|
|
+ self._log("\n=== 步骤7: 生成K线图可视化 ===")
|
|
|
+
|
|
|
+ max_charts = min(len(pattern_results), self.config['output_settings']['max_chart_count'])
|
|
|
+
|
|
|
+ for i, pattern in enumerate(pattern_results[:max_charts]):
|
|
|
+ code = pattern['品种代码']
|
|
|
+ pattern_date = pattern['日期']
|
|
|
+ pattern_idx = pattern['数据索引']
|
|
|
+
|
|
|
+ if code not in weekly_data:
|
|
|
+ continue
|
|
|
+
|
|
|
+ data = weekly_data[code]['data']
|
|
|
+
|
|
|
+ # 准备绘图数据（形态前后各5周）
|
|
|
+ start_idx = max(0, pattern_idx - 5)
|
|
|
+ end_idx = min(len(data), pattern_idx + 11)
|
|
|
+ plot_data = data.iloc[start_idx:end_idx]
|
|
|
+
|
|
|
+ # 创建K线图
|
|
|
+ fig, ax = plt.subplots(figsize=(12, 8))
|
|
|
+
|
|
|
+ # 绘制K线
|
|
|
+ for j, (date, row) in enumerate(plot_data.iterrows()):
|
|
|
+ open_price = row['open']
|
|
|
+ close_price = row['close']
|
|
|
+ high_price = row['high']
|
|
|
+ low_price = row['low']
|
|
|
+
|
|
|
+ # K线颜色
|
|
|
+ color = 'red' if close_price > open_price else 'green'
|
|
|
+
|
|
|
+ # 绘制影线
|
|
|
+ ax.plot([j, j], [low_price, high_price], color='black', linewidth=1)
|
|
|
+
|
|
|
+ # 绘制实体
|
|
|
+ body_height = abs(close_price - open_price)
|
|
|
+ if close_price > open_price:
|
|
|
+ rect = patches.Rectangle((j-0.3, open_price), 0.6, body_height,
|
|
|
+ linewidth=1, edgecolor='red', facecolor='red', alpha=0.7)
|
|
|
+ else:
|
|
|
+ rect = patches.Rectangle((j-0.3, close_price), 0.6, body_height,
|
|
|
+ linewidth=1, edgecolor='green', facecolor='green', alpha=0.7)
|
|
|
+ ax.add_patch(rect)
|
|
|
+
|
|
|
+ # 标记目标形态
|
|
|
+ if date == pattern_date:
|
|
|
+ ax.scatter(j, high_price + (high_price - low_price) * 0.1,
|
|
|
+ color='yellow', s=100, marker='*', zorder=5)
|
|
|
+ ax.text(j, high_price + (high_price - low_price) * 0.15,
|
|
|
+ '目标形态', ha='center', fontsize=10, color='red', fontweight='bold')
|
|
|
+
|
|
|
+ # 设置图表
|
|
|
+ ax.set_title(f'{code} 周线K线图 - 目标形态分析\n形态日期: {pattern_date.strftime("%Y-%m-%d")}',
|
|
|
+ fontsize=14, fontweight='bold')
|
|
|
+ ax.set_xlabel('时间', fontsize=12)
|
|
|
+ ax.set_ylabel('价格', fontsize=12)
|
|
|
+ ax.grid(True, alpha=0.3)
|
|
|
+
|
|
|
+ # 设置x轴标签
|
|
|
+ dates = [date.strftime('%m-%d') for date in plot_data.index]
|
|
|
+ ax.set_xticks(range(len(dates)))
|
|
|
+ ax.set_xticklabels(dates, rotation=45)
|
|
|
+
|
|
|
+ # 显示图表
|
|
|
+ plt.tight_layout()
|
|
|
+ plt.show() # 直接在界面中显示图表
|
|
|
+ self._log(f"显示图表: {code} {pattern_date.strftime('%Y-%m-%d')} K线形态分析")
|
|
|
+
|
|
|
+ def generate_summary_report(self, decline_futures, pattern_results, performance_results):
|
|
|
+ """生成整体规律总结报告"""
|
|
|
+ self._log("\n=== 步骤8: 生成统计报告 ===")
|
|
|
+
|
|
|
+ # 基础统计
|
|
|
+ total_decline_futures = len(decline_futures)
|
|
|
+ total_patterns = len(pattern_results)
|
|
|
+ total_analyzed = len(performance_results)
|
|
|
+
|
|
|
+ print(f"\n=== 期货技术形态规律分析报告 ===")
|
|
|
+ print(f"分析期间: {self.start_date.strftime('%Y-%m-%d')} 至 {self.end_date.strftime('%Y-%m-%d')}")
|
|
|
+ print(f"总分析期货品种数: {len(self.future_codes)}")
|
|
|
+ print(f"大幅下跌品种数: {total_decline_futures}")
|
|
|
+ print(f"识别出的目标形态数: {total_patterns}")
|
|
|
+ print(f"完成后续走势分析数: {total_analyzed}")
|
|
|
+
|
|
|
+ if total_analyzed == 0:
|
|
|
+ print("无有效数据进行统计分析")
|
|
|
+ return {}
|
|
|
+
|
|
|
+ # 收益率统计
|
|
|
+ print(f"\n=== 后续走势统计分析 ===")
|
|
|
+
|
|
|
+ # 提取所有统计指标
|
|
|
+ avg_weekly_changes = [r['统计_平均周涨跌幅'] for r in performance_results if '统计_平均周涨跌幅' in r]
|
|
|
+ max_weekly_gains = [r['统计_最大周涨幅'] for r in performance_results if '统计_最大周涨幅' in r]
|
|
|
+ max_weekly_losses = [r['统计_最大周跌幅'] for r in performance_results if '统计_最大周跌幅' in r]
|
|
|
+ avg_max_declines = [r['统计_平均最大跌幅'] for r in performance_results if '统计_平均最大跌幅' in r]
|
|
|
+
|
|
|
+ if avg_weekly_changes:
|
|
|
+ print(f"平均周涨跌幅统计:")
|
|
|
+ print(f" 平均值: {np.mean(avg_weekly_changes):.2f}%")
|
|
|
+ print(f" 中位数: {np.median(avg_weekly_changes):.2f}%")
|
|
|
+ print(f" 最大值: {max(avg_weekly_changes):.2f}%")
|
|
|
+ print(f" 最小值: {min(avg_weekly_changes):.2f}%")
|
|
|
+
|
|
|
+ positive_weeks = sum(1 for r in avg_weekly_changes if r > 0)
|
|
|
+ print(f" 正收益形态比例: {positive_weeks}/{len(avg_weekly_changes)} ({positive_weeks/len(avg_weekly_changes)*100:.1f}%)")
|
|
|
+
|
|
|
+ if max_weekly_gains:
|
|
|
+ print(f"\n最大周涨幅统计:")
|
|
|
+ print(f" 平均值: {np.mean(max_weekly_gains):.2f}%")
|
|
|
+ print(f" 中位数: {np.median(max_weekly_gains):.2f}%")
|
|
|
+ print(f" 最大值: {max(max_weekly_gains):.2f}%")
|
|
|
+ print(f" 最小值: {min(max_weekly_gains):.2f}%")
|
|
|
+
|
|
|
+ if max_weekly_losses:
|
|
|
+ print(f"\n最大周跌幅统计:")
|
|
|
+ print(f" 平均值: {np.mean(max_weekly_losses):.2f}%")
|
|
|
+ print(f" 中位数: {np.median(max_weekly_losses):.2f}%")
|
|
|
+ print(f" 最大值: {max(max_weekly_losses):.2f}%")
|
|
|
+ print(f" 最小值: {min(max_weekly_losses):.2f}%")
|
|
|
+
|
|
|
+ if avg_max_declines:
|
|
|
+ print(f"\n平均最大跌幅统计:")
|
|
|
+ print(f" 平均值: {np.mean(avg_max_declines):.2f}%")
|
|
|
+ print(f" 中位数: {np.median(avg_max_declines):.2f}%")
|
|
|
+ print(f" 最大值: {max(avg_max_declines):.2f}%")
|
|
|
+ print(f" 最小值: {min(avg_max_declines):.2f}%")
|
|
|
+
|
|
|
+ # 成功率统计 - 基于平均周涨跌幅是否为正
|
|
|
+ positive_performance_count = sum(1 for r in avg_weekly_changes if r > 0) if avg_weekly_changes else 0
|
|
|
+ success_rate = positive_performance_count / len(avg_weekly_changes) * 100 if avg_weekly_changes else 0
|
|
|
+
|
|
|
+ # 基于最大周涨幅的成功率统计
|
|
|
+ success_threshold = self.config['analysis_params']['success_threshold'] * 100
|
|
|
+ high_gain_count = sum(1 for r in max_weekly_gains if r > success_threshold) if max_weekly_gains else 0
|
|
|
+ high_gain_rate = high_gain_count / len(max_weekly_gains) * 100 if max_weekly_gains else 0
|
|
|
+
|
|
|
+ print(f"\n=== 交易成功率分析 ===")
|
|
|
+ print(f"正收益形态数量: {positive_performance_count}/{len(avg_weekly_changes) if avg_weekly_changes else 0}")
|
|
|
+ print(f"正收益成功率: {success_rate:.1f}%")
|
|
|
+ print(f"")
|
|
|
+ print(f"高收益标准: 最大周涨幅 > {success_threshold}%")
|
|
|
+ print(f"高收益形态数量: {high_gain_count}/{len(max_weekly_gains) if max_weekly_gains else 0}")
|
|
|
+ print(f"高收益成功率: {high_gain_rate:.1f}%")
|
|
|
+
|
|
|
+ return {
|
|
|
+ 'total_futures': len(self.future_codes),
|
|
|
+ 'decline_futures': total_decline_futures,
|
|
|
+ 'pattern_count': total_patterns,
|
|
|
+ 'analyzed_count': total_analyzed,
|
|
|
+ 'avg_weekly_change': np.mean(avg_weekly_changes) if avg_weekly_changes else 0,
|
|
|
+ 'avg_max_gain': np.mean(max_weekly_gains) if max_weekly_gains else 0,
|
|
|
+ 'avg_max_loss': np.mean(max_weekly_losses) if max_weekly_losses else 0,
|
|
|
+ 'avg_max_decline': np.mean(avg_max_declines) if avg_max_declines else 0,
|
|
|
+ 'success_rate': success_rate,
|
|
|
+ 'high_gain_rate': high_gain_rate,
|
|
|
+ 'positive_count': positive_performance_count,
|
|
|
+ 'high_gain_count': high_gain_count
|
|
|
+ }
|
|
|
+
|
|
|
+ def save_results(self, decline_futures, pattern_results, performance_results, summary_stats):
|
|
|
+ """保存结果到CSV文件"""
|
|
|
+ self._log("\n=== 保存分析结果 ===")
|
|
|
+
|
|
|
+ timestamp = datetime.datetime.now().strftime('%Y%m%d_%H%M%S')
|
|
|
+ encoding = self.config['output_settings']['csv_encoding']
|
|
|
+
|
|
|
+ # 保存大幅下跌期货品种
|
|
|
+ if decline_futures:
|
|
|
+ decline_df = pd.DataFrame(decline_futures)
|
|
|
+ decline_file = f'future_major_decline_{timestamp}.csv'
|
|
|
+ decline_df.to_csv(decline_file, index=False, encoding=encoding)
|
|
|
+ self._log(f"大幅下跌期货保存至: {decline_file}")
|
|
|
+
|
|
|
+ # 保存识别的K线形态
|
|
|
+ if pattern_results:
|
|
|
+ pattern_df = pd.DataFrame(pattern_results)
|
|
|
+ pattern_file = f'future_kline_patterns_{timestamp}.csv'
|
|
|
+ pattern_df.to_csv(pattern_file, index=False, encoding=encoding)
|
|
|
+ self._log(f"K线形态数据保存至: {pattern_file}")
|
|
|
+
|
|
|
+ # 保存后续走势分析
|
|
|
+ if performance_results:
|
|
|
+ performance_df = pd.DataFrame(performance_results)
|
|
|
+ performance_file = f'future_pattern_performance_{timestamp}.csv'
|
|
|
+ performance_df.to_csv(performance_file, index=False, encoding=encoding)
|
|
|
+ self._log(f"后续走势分析保存至: {performance_file}")
|
|
|
+
|
|
|
+ # 保存汇总报告
|
|
|
+ if summary_stats:
|
|
|
+ summary_df = pd.DataFrame([summary_stats])
|
|
|
+ summary_file = f'future_pattern_summary_{timestamp}.csv'
|
|
|
+ summary_df.to_csv(summary_file, index=False, encoding=encoding)
|
|
|
+ self._log(f"汇总报告保存至: {summary_file}")
|
|
|
+
|
|
|
+ def run_analysis(self, mode=1, manual_config=None):
|
|
|
+ """
|
|
|
+ 统一的分析运行入口
|
|
|
+
|
|
|
+ 参数:
|
|
|
+ - mode: 执行模式
|
|
|
+ 1 = 原始/自动模式：自动搜索符合条件的技术形态
|
|
|
+ 2 = 验证模式：验证指定日期是否符合模式条件
|
|
|
+ - manual_config: 手动配置字典，仅在mode=2时使用
|
|
|
+
|
|
|
+ 返回:
|
|
|
+ - dict: 分析结果
|
|
|
+ """
|
|
|
+
|
|
|
+ if mode == 1:
|
|
|
+ return self._run_original_mode()
|
|
|
+ elif mode == 2:
|
|
|
+ if manual_config is None:
|
|
|
+ raise ValueError("验证模式(mode=2)需要提供manual_config参数")
|
|
|
+ return self._run_validation_mode(manual_config)
|
|
|
+ else:
|
|
|
+ raise ValueError("mode参数必须为1(原始模式)或2(验证模式)")
|
|
|
+
|
|
|
+ def _run_original_mode(self):
|
|
|
+ """原始/自动模式：自动搜索符合条件的技术形态"""
|
|
|
+ print("开始期货技术形态规律分析... [原始模式]")
|
|
|
+
|
|
|
+ # 构建期货品种基础数据
|
|
|
+ self.build_future_codes()
|
|
|
+
|
|
|
+ # 筛选大幅下跌的期货品种
|
|
|
+ decline_futures = self.find_major_decline_futures()
|
|
|
+
|
|
|
+ if not decline_futures:
|
|
|
+ print("未找到符合条件的大幅下跌期货品种")
|
|
|
+ return {}
|
|
|
+
|
|
|
+ # 创建8888加权合约映射
|
|
|
+ contract_mapping = self.get_simplified_contract_mapping(decline_futures)
|
|
|
+
|
|
|
+ if not contract_mapping:
|
|
|
+ print("未能创建8888合约映射")
|
|
|
+ return {}
|
|
|
+ else:
|
|
|
+ print(f"创建8888合约映射成功，共{len(contract_mapping)}个合约")
|
|
|
+ for contract_code, time_range in contract_mapping.items():
|
|
|
+ start_date, end_date = time_range
|
|
|
+ print(f" {contract_code}: {start_date.strftime('%Y-%m-%d')} 至 {end_date.strftime('%Y-%m-%d')}")
|
|
|
+
|
|
|
+ # 获取周线K线数据
|
|
|
+ weekly_data = self.get_weekly_kline_data(contract_mapping)
|
|
|
+
|
|
|
+ if not weekly_data:
|
|
|
+ print("未能获取周线数据")
|
|
|
+ return {}
|
|
|
+
|
|
|
+ # 识别目标K线形态
|
|
|
+ pattern_results = self.identify_target_pattern(weekly_data, decline_futures)
|
|
|
+
|
|
|
+ if not pattern_results:
|
|
|
+ print("未识别出目标K线形态")
|
|
|
+ return {}
|
|
|
+
|
|
|
+ # 分析后续走势
|
|
|
+ performance_results = self.analyze_future_performance(pattern_results, weekly_data)
|
|
|
+
|
|
|
+ # 创建可视化
|
|
|
+ self.create_visualizations(pattern_results, weekly_data)
|
|
|
+
|
|
|
+ # 生成汇总报告
|
|
|
+ summary_stats = self.generate_summary_report(decline_futures, pattern_results, performance_results)
|
|
|
+
|
|
|
+ print(f"\n=== 分析完成 ===")
|
|
|
+ print(f"共分析了{len(decline_futures)}个大幅下跌期货品种")
|
|
|
+ print(f"识别出{len(pattern_results)}个目标K线形态")
|
|
|
+ print(f"完成{len(performance_results)}个形态的后续走势分析")
|
|
|
+
|
|
|
+ return {
|
|
|
+ 'mode': 1,
|
|
|
+ 'decline_futures': decline_futures,
|
|
|
+ 'pattern_results': pattern_results,
|
|
|
+ 'performance_results': performance_results,
|
|
|
+ 'summary_stats': summary_stats
|
|
|
+ }
|
|
|
+
|
|
|
+ def _run_validation_mode(self, manual_config):
|
|
|
+ """验证模式：验证指定日期是否符合模式条件"""
|
|
|
+ print("开始期货技术形态验证分析... [验证模式]")
|
|
|
+
|
|
|
+ # 构建期货品种基础数据（仅构建需要的品种）
|
|
|
+ self.build_future_codes()
|
|
|
+
|
|
|
+ # 从配置中提取需要验证的品种代码
|
|
|
+ config_codes = list(manual_config.keys())
|
|
|
+ print(f"配置的验证目标: {config_codes}")
|
|
|
+
|
|
|
+ # 为每个配置的品种创建数据获取范围
|
|
|
+ sample_futures = []
|
|
|
+
|
|
|
+ for code in config_codes:
|
|
|
+ if code in self.future_codes:
|
|
|
+ contract_8888 = self.future_codes[code]
|
|
|
+ date_list = manual_config[code]
|
|
|
+
|
|
|
+ # 基于配置的日期计算数据获取范围
|
|
|
+ config_dates = [pd.to_datetime(date_str) for date_str in date_list]
|
|
|
+ earliest_date = min(config_dates)
|
|
|
+ latest_date = max(config_dates)
|
|
|
+
|
|
|
+ # 为了确保有足够的数据进行验证，在日期范围前后各扩展一些时间
|
|
|
+ buffer_days = 90 # 3个月缓冲
|
|
|
+ start_date = earliest_date - pd.Timedelta(days=buffer_days)
|
|
|
+ end_date = latest_date + pd.Timedelta(days=buffer_days)
|
|
|
+
|
|
|
+ sample_futures.append({
|
|
|
+ '品种代码': code,
|
|
|
+ '合约代码': contract_8888,
|
|
|
+ '高点日期': start_date,
|
|
|
+ '低点日期': end_date
|
|
|
+ })
|
|
|
+
|
|
|
+ print(f" ✓ {code} ({contract_8888}): {start_date.strftime('%Y-%m-%d')} 至 {end_date.strftime('%Y-%m-%d')}")
|
|
|
+ else:
|
|
|
+ print(f" ⚠️ 未找到品种 '{code}' 对应的合约代码")
|
|
|
+
|
|
|
+ print(f"实际处理的品种数: {len(sample_futures)}")
|
|
|
+
|
|
|
+ if not sample_futures:
|
|
|
+ print("❌ 配置中的品种代码都无法找到对应的合约")
|
|
|
+ return {}
|
|
|
+
|
|
|
+ # 创建简化的合约映射
|
|
|
+ contract_mapping = self.get_simplified_contract_mapping(sample_futures)
|
|
|
+
|
|
|
+ if not contract_mapping:
|
|
|
+ print("❌ 无法创建合约映射")
|
|
|
+ return {}
|
|
|
+
|
|
|
+ # 获取周线数据
|
|
|
+ weekly_data = self.get_weekly_kline_data(contract_mapping)
|
|
|
+
|
|
|
+ if not weekly_data:
|
|
|
+ print("❌ 无法获取周线数据")
|
|
|
+ return {}
|
|
|
+
|
|
|
+ print(f"✓ 成功获取 {len(weekly_data)} 个品种的数据")
|
|
|
+
|
|
|
+ # 执行手动验证
|
|
|
+ validation_results = self.validate_manual_patterns(manual_config, weekly_data)
|
|
|
+ validation_results['mode'] = 2
|
|
|
+
|
|
|
+ return validation_results
|
|
|
+
|
|
|
+# =====================================================================================
|
|
|
+# 第三部分：主程序运行函数
|
|
|
+# 提供统一的程序入口和运行选项
|
|
|
+# =====================================================================================
|
|
|
+
|
|
|
+def run_pattern_analysis(mode=1, manual_config=None):
|
|
|
+ """
|
|
|
+ 统一的期货技术形态分析入口
|
|
|
+
|
|
|
+ 参数:
|
|
|
+ - mode: 执行模式
|
|
|
+ 1 = 原始/自动模式：自动搜索符合条件的技术形态
|
|
|
+ 2 = 验证模式：验证指定日期是否符合模式条件
|
|
|
+ - manual_config: 手动配置字典，仅在mode=2时使用
|
|
|
+
|
|
|
+ 返回:
|
|
|
+ - dict: 分析结果
|
|
|
+ """
|
|
|
+ print("期货技术形态规律分析工具")
|
|
|
+ print("="*60)
|
|
|
+
|
|
|
+ # 打印配置信息
|
|
|
+ AnalysisConfig.print_config()
|
|
|
+ print("="*60)
|
|
|
+
|
|
|
+ mode_name = "原始模式" if mode == 1 else "验证模式"
|
|
|
+ print(f"开始分析... [{mode_name}]")
|
|
|
+
|
|
|
+ # 创建分析器并运行
|
|
|
+ analyzer = FuturePatternAnalyzer(
|
|
|
+ AnalysisConfig.START_DATE,
|
|
|
+ AnalysisConfig.END_DATE
|
|
|
+ )
|
|
|
+
|
|
|
+ results = analyzer.run_analysis(mode=mode, manual_config=manual_config)
|
|
|
+ return results
|
|
|
+
|
|
|
+# =====================================================================================
|
|
|
+# 第四部分：辅助工具函数
|
|
|
+# 用于处理合约切换日志和数据导出
|
|
|
+# =====================================================================================
|
|
|
+
|
|
|
+def parse_contract_switch_log_from_text(log_text):
|
|
|
+ """
|
|
|
+ 从文本格式的合约切换日志中解析合约信息
|
|
|
+
|
|
|
+ 参数:
|
|
|
+ - log_text: 日志文本，格式示例：
|
|
|
+ "A2203.XDCE：2022年1月26日至2022年2月22日
|
|
|
+ A2207.XDCE：2022年2月23日至2022年6月21日"
|
|
|
+
|
|
|
+ 返回:
|
|
|
+ - list: 标准化的合约切换日志
|
|
|
+ """
|
|
|
+ import re
|
|
|
+
|
|
|
+ contracts = []
|
|
|
+ lines = log_text.strip().split('\n')
|
|
|
+
|
|
|
+ for line in lines:
|
|
|
+ line = line.strip()
|
|
|
+ if not line:
|
|
|
+ continue
|
|
|
+
|
|
|
+ # 正则匹配合约代码和日期范围
|
|
|
+ pattern = r'([A-Z0-9]+\.[A-Z]+)：(\d{4})年(\d{1,2})月(\d{1,2})日至(\d{4})年(\d{1,2})月(\d{1,2})日'
|
|
|
+ match = re.search(pattern, line)
|
|
|
+
|
|
|
+ if match:
|
|
|
+ contract_code = match.group(1)
|
|
|
+ start_year = int(match.group(2))
|
|
|
+ start_month = int(match.group(3))
|
|
|
+ start_day = int(match.group(4))
|
|
|
+ end_year = int(match.group(5))
|
|
|
+ end_month = int(match.group(6))
|
|
|
+ end_day = int(match.group(7))
|
|
|
+
|
|
|
+ contracts.append({
|
|
|
+ 'contract': contract_code,
|
|
|
+ 'start_date': f'{start_year}-{start_month:02d}-{start_day:02d}',
|
|
|
+ 'end_date': f'{end_year}-{end_month:02d}-{end_day:02d}'
|
|
|
+ })
|
|
|
+
|
|
|
+ return contracts
|
|
|
+
|
|
|
+def export_contract_weekly_data(weekly_data_dict, filename_prefix='contract_weekly_data'):
|
|
|
+ """
|
|
|
+ 导出合约周线数据到CSV文件
|
|
|
+
|
|
|
+ 参数:
|
|
|
+ - weekly_data_dict: get_weekly_data_by_contract_log返回的数据字典
|
|
|
+ - filename_prefix: 文件名前缀
|
|
|
+
|
|
|
+ 注意：导出的数据是直接从get_bars API获取的原生周线数据，
|
|
|
+ 相比日线转换的数据具有更高的精度和准确性
|
|
|
+ """
|
|
|
+ import datetime
|
|
|
+
|
|
|
+ timestamp = datetime.datetime.now().strftime('%Y%m%d_%H%M%S')
|
|
|
+
|
|
|
+ for contract_code, contract_data in weekly_data_dict.items():
|
|
|
+ # 清理文件名中的特殊字符
|
|
|
+ safe_contract_name = contract_code.replace('.', '_')
|
|
|
+ filename = f'{filename_prefix}_{safe_contract_name}_{timestamp}.csv'
|
|
|
+
|
|
|
+ # 准备数据
|
|
|
+ df = contract_data['data'].copy()
|
|
|
+ df['contract'] = contract_code
|
|
|
+ df['original_start'] = contract_data['original_start'].strftime('%Y-%m-%d')
|
|
|
+ df['original_end'] = contract_data['original_end'].strftime('%Y-%m-%d')
|
|
|
+
|
|
|
+ # 保存到CSV
|
|
|
+ df.to_csv(filename, encoding='utf-8-sig')
|
|
|
+ print(f"合约 {contract_code} 的周线数据已保存至: {filename}")
|
|
|
+
|
|
|
+def run_manual_pattern_validation(manual_config, analyzer=None):
|
|
|
+ """
|
|
|
+ 运行手动模式验证的便捷函数
|
|
|
+
|
|
|
+ 此函数只会处理manual_config中明确指定的品种和时间范围，
|
|
|
+ 不会查询额外的品种数据。现在使用统一的分析框架。
|
|
|
+
|
|
|
+ 参数:
|
|
|
+ - manual_config: 手动配置字典，格式：
|
|
|
+ {'A':['2023-01-06','2023-06-09'], 'B':['2022-07-22','2024-08-23']}
|
|
|
+ 键: 期货品种代码（如'A'、'B'、'RB'等）
|
|
|
+ 值: 需要验证的日期列表（'YYYY-MM-DD'格式）
|
|
|
+ - analyzer: 可选的分析器实例，如果为None则创建新实例
|
|
|
+
|
|
|
+ 返回:
|
|
|
+ - dict: 验证结果，包含成功和失败的详细信息
|
|
|
+
|
|
|
+ 注意:
|
|
|
+ - 系统只会获取配置中指定品种的数据
|
|
|
+ - 数据获取范围基于每个品种配置的日期范围自动计算
|
|
|
+ - 不会处理未在配置中明确指定的品种
|
|
|
+ - 现在使用统一的核心验证逻辑，确保与原始模式的一致性
|
|
|
+ """
|
|
|
+ print("\n=== 手动模式验证工具 ===")
|
|
|
+ print("此功能用于验证指定日期是否符合预定义的技术形态模式")
|
|
|
+ print("使用统一的验证逻辑，与原始模式保持完全一致")
|
|
|
+
|
|
|
+ # 创建分析器实例（如果没有提供）
|
|
|
+ if analyzer is None:
|
|
|
+ analyzer = FuturePatternAnalyzer()
|
|
|
+
|
|
|
+ # 使用统一的分析入口 - 验证模式
|
|
|
+ validation_results = analyzer.run_analysis(mode=2, manual_config=manual_config)
|
|
|
+
|
|
|
+ return validation_results
|
|
|
+
|
|
|
+def demo_manual_validation():
|
|
|
+ """演示手动模式验证功能"""
|
|
|
+ print("\n=== 手动模式验证演示 ===")
|
|
|
+
|
|
|
+ # 示例配置
|
|
|
+ demo_config = {
|
|
|
+ 'A': ['2023-01-06', '2023-06-09', '2024-07-26'],
|
|
|
+ 'B': ['2022-07-22', '2024-08-23', '2025-01-03']
|
|
|
+ }
|
|
|
+
|
|
|
+ print("使用演示配置:")
|
|
|
+ for pattern_type, dates in demo_config.items():
|
|
|
+ print(f" 模式 '{pattern_type}': {len(dates)} 个日期")
|
|
|
+ for date in dates:
|
|
|
+ print(f" - {date}")
|
|
|
+
|
|
|
+ try:
|
|
|
+ # 运行验证
|
|
|
+ results = run_manual_pattern_validation(demo_config)
|
|
|
+
|
|
|
+ if results:
|
|
|
+ print("\n✅ 验证完成")
|
|
|
+ return results
|
|
|
+ else:
|
|
|
+ print("\n⚠️ 验证未能完成")
|
|
|
+ return {}
|
|
|
+
|
|
|
+ except Exception as e:
|
|
|
+ print(f"\n❌ 验证过程中出错: {str(e)}")
|
|
|
+ import traceback
|
|
|
+ traceback.print_exc()
|
|
|
+ return {}
|
|
|
+
|
|
|
+def get_contract_weekly_data(contract_switch_log, extend_weeks=10, export_data=False):
|
|
|
+ """
|
|
|
+ 基于合约切换日志获取周线数据的正式方法
|
|
|
+
|
|
|
+ 参数:
|
|
|
+ - contract_switch_log: 合约切换日志，可以是文本格式或列表格式
|
|
|
+ - extend_weeks: 延长周数，默认10周
|
|
|
+ - export_data: 是否导出数据到CSV文件，默认False
|
|
|
+
|
|
|
+ 返回:
|
|
|
+ - dict: 每个合约的周线数据
|
|
|
+ """
|
|
|
+ print("\n=== 基于合约切换日志获取周线数据 ===")
|
|
|
+
|
|
|
+ # 如果是文本格式，先解析
|
|
|
+ if isinstance(contract_switch_log, str):
|
|
|
+ print("解析合约切换日志:")
|
|
|
+ contracts = parse_contract_switch_log_from_text(contract_switch_log)
|
|
|
+ for contract in contracts:
|
|
|
+ print(f" {contract['contract']}: {contract['start_date']} 至 {contract['end_date']}")
|
|
|
+ else:
|
|
|
+ contracts = contract_switch_log
|
|
|
+ print(f"处理{len(contracts)}个合约的切换日志")
|
|
|
+
|
|
|
+ print(f"\n创建分析器并获取周线数据 (延长{extend_weeks}周):")
|
|
|
+ analyzer = FuturePatternAnalyzer()
|
|
|
+
|
|
|
+ try:
|
|
|
+ # 获取周线数据
|
|
|
+ weekly_data = analyzer.get_weekly_data_by_contract_log(contracts, extend_weeks=extend_weeks)
|
|
|
+
|
|
|
+ if weekly_data:
|
|
|
+ print(" ✓ 获取成功！")
|
|
|
+ for contract_code, data_info in weekly_data.items():
|
|
|
+ print(f" {contract_code}: {data_info['total_weeks']}周数据")
|
|
|
+ print(f" 原始范围: {data_info['original_start'].strftime('%Y-%m-%d')} 至 {data_info['original_end'].strftime('%Y-%m-%d')}")
|
|
|
+ print(f" 延长范围: {data_info['extended_start'].strftime('%Y-%m-%d')} 至 {data_info['extended_end'].strftime('%Y-%m-%d')}")
|
|
|
+
|
|
|
+ # 导出数据
|
|
|
+ if export_data:
|
|
|
+ export_contract_weekly_data(weekly_data)
|
|
|
+
|
|
|
+ return weekly_data
|
|
|
+ else:
|
|
|
+ print(" ⚠️ 未获取到数据")
|
|
|
+ return {}
|
|
|
+
|
|
|
+ except Exception as e:
|
|
|
+ print(f" ❌ 获取失败: {str(e)}")
|
|
|
+ import traceback
|
|
|
+ traceback.print_exc()
|
|
|
+ return {}
|
|
|
+
|
|
|
+# =====================================================================================
|
|
|
+# 第五部分：程序主入口 - 自动运行分析
|
|
|
+# =====================================================================================
|
|
|
+
|
|
|
+print("期货市场技术形态规律分析工具")
|
|
|
+print("本工具专门研究期货品种在大幅下跌后出现特定K线形态的规律")
|
|
|
+print("适用于聚宽在线研究平台")
|
|
|
+print("="*80)
|
|
|
+
|
|
|
+# =====================================================================================
|
|
|
+# 基于合约切换日志的周线数据获取功能使用说明
|
|
|
+# =====================================================================================
|
|
|
+#
|
|
|
+# 使用示例1：处理文本格式的合约切换日志
|
|
|
+# sample_log = """
|
|
|
+# A2203.XDCE：2022年1月26日至2022年2月22日
|
|
|
+# A2207.XDCE：2022年2月23日至2022年6月21日
|
|
|
+# A2209.XDCE：2022年6月22日至2022年8月30日
|
|
|
+# """
|
|
|
+# weekly_data = get_contract_weekly_data(sample_log, extend_weeks=10, export_data=True)
|
|
|
+#
|
|
|
+# 使用示例2：处理列表格式的合约切换日志
|
|
|
+# contract_list = [
|
|
|
+# {'contract': 'A2203.XDCE', 'start_date': '2022-01-26', 'end_date': '2022-02-22'},
|
|
|
+# {'contract': 'A2207.XDCE', 'start_date': '2022-02-23', 'end_date': '2022-06-21'}
|
|
|
+# ]
|
|
|
+# weekly_data = get_contract_weekly_data(contract_list, extend_weeks=15, export_data=False)
|
|
|
+#
|
|
|
+# 新增功能1：模式识别失败时的完整K线图显示
|
|
|
+# 当identify_target_pattern函数未识别到任何目标形态时，系统会自动：
|
|
|
+# 1. 显示所有可用品种的完整烛台图
|
|
|
+# 2. 包含OHLC数据的专业K线图表现
|
|
|
+# 3. 添加统计信息（最高价、最低价、总涨跌幅等）
|
|
|
+# 4. 直接在界面中显示，不保存到本地文件
|
|
|
+#
|
|
|
+# 新增功能2：手动模式验证系统
|
|
|
+# 使用示例3：手动验证指定日期的模式匹配
|
|
|
+# 现在使用统一的入口函数，通过mode参数控制执行模式
|
|
|
+manual_config = {
|
|
|
+ 'A': ['2023-01-06', '2023-06-09', '2024-07-26'],
|
|
|
+ 'B': ['2022-07-22', '2024-08-23', '2025-01-03']
|
|
|
+}
|
|
|
+
|
|
|
+# 方法1: 使用统一入口函数
|
|
|
+validation_results = run_pattern_analysis(mode=2, manual_config=manual_config)
|
|
|
+
|
|
|
+# 方法2: 或者使用便捷函数（内部调用统一入口）
|
|
|
+# validation_results = run_manual_pattern_validation(manual_config)
|
|
|
+# 或运行演示: demo_manual_validation()
|
|
|
+#
|
|
|
+# 重构后的统一模式控制系统：
|
|
|
+# 1. mode=1（原始模式）：自动搜索符合条件的技术形态，遍历所有期货品种
|
|
|
+# 2. mode=2（验证模式）：验证指定日期是否符合模式条件，只处理配置品种
|
|
|
+# 3. 统一的核心验证逻辑：_core_pattern_validation()消除了代码重复
|
|
|
+# 4. 独立的执行路径：两种模式不再共享步骤1，完全独立执行
|
|
|
+# 5. 一致的验证标准：两种模式使用完全相同的模式识别条件和计算方法
|
|
|
+#
|
|
|
+# 数据获取特点：
|
|
|
+# 1. 直接使用get_bars API获取原生周线数据，避免日线转换的精度损失
|
|
|
+# 2. 自动延长时间范围（每个合约前后各延长指定周数）
|
|
|
+# 3. 确保获取充足的历史数据（最少60周作为背景）
|
|
|
+# 4. 支持数据导出到CSV文件便于后续分析
|
|
|
+# 5. 简化结构：直接使用8888加权合约，无需处理主力合约切换
|
|
|
+#
|
|
|
+# 简化后的工作流程：
|
|
|
+# 1. 识别符合条件的大幅下跌期货品种
|
|
|
+# 2. 为每个品种创建8888合约映射（单级字典结构）
|
|
|
+# 3. 直接从8888合约获取周线数据，避免主力合约复杂性
|
|
|
+# 4. 进行技术形态识别和后续走势分析
|
|
|
+# 5. 当未识别到目标形态时，自动显示完整的烛台图（K线图）
|
|
|
+# =====================================================================================
|
|
|
+
|
|
|
+# 直接运行分析
|
|
|
+# print("\n开始运行完整分析...")
|
|
|
+# try:
|
|
|
+# results = run_pattern_analysis()
|
|
|
+# except Exception as e:
|
|
|
+# print(f"运行过程中出现错误: {str(e)}")
|
|
|
+# import traceback
|
|
|
+# traceback.print_exc()
|
|
|
+# results = None
|
|
|
+
|
|
|
+# if results:
|
|
|
+# print("\n✅ 分析完成！")
|
|
|
+# if isinstance(results, dict) and 'summary_stats' in results:
|
|
|
+# summary = results['summary_stats']
|
|
|
+# print(f"\n📊 快速统计:")
|
|
|
+# print(f" - 识别到技术形态: {summary.get('pattern_count', 0)}个")
|
|
|
+# print(f" - 正收益成功率: {summary.get('success_rate', 0):.1f}%")
|
|
|
+# print(f" - 高收益成功率: {summary.get('high_gain_rate', 0):.1f}%")
|
|
|
+# print(f" - 平均周涨跌幅: {summary.get('avg_weekly_change', 0):+.2f}%")
|
|
|
+# else:
|
|
|
+# print("\n❌ 分析未能完成，请检查错误信息。")
|
maxfeng