huice/analyze_large_orders.py

#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""
AU2512期货大单分析工具 (买1量和卖1量整合分析)

这个脚本专门分析买1量和卖1量大于阈值的数据，并计算这些时点后100和200个成交笔数内的价格范围。
买1量大通常表示存在大量买盘挂单，可能是价格上涨的信号。
卖1量大通常表示存在大量卖盘挂单，可能是价格下跌的信号。

使用方法:
python analyze_large_orders.py [data_file]

如果不指定文件，默认分析 data/au2512_20251013.parquet
输出目录: large_orders/
"""

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import warnings
import os
import sys
import argparse
from pathlib import Path

warnings.filterwarnings('ignore')

class LargeOrdersAnalyzer:
    """大单分析器 (买1量 + 卖1量)"""

    def __init__(self, data_file=None, output_dir="large_orders"):
        """
        初始化分析器

        Args:
            data_file (str): 数据文件路径，默认为 data/au2512_20251013.parquet
            output_dir (str): 输出目录，默认为 large_orders
        """
        self.data_file = data_file or "data/au2512_20251013.parquet"
        self.output_dir = output_dir

        # 确保输出目录存在
        os.makedirs(self.output_dir, exist_ok=True)

        self.df = None
        self.buy1_large_orders = None
        self.sell1_large_orders = None
        self.buy1_results_100 = None
        self.buy1_results_200 = None
        self.sell1_results_100 = None
        self.sell1_results_200 = None
        self.setup_chinese_font()

    def setup_chinese_font(self):
        """设置中文字体支持"""
        try:
            # 尝试不同的中文字体
            chinese_fonts = ['Microsoft YaHei', 'SimHei', 'SimSun', 'KaiTi', 'FangSong']

            for font in chinese_fonts:
                try:
                    plt.rcParams['font.sans-serif'] = [font]
                    plt.rcParams['axes.unicode_minus'] = False
                    # 测试字体是否可用
                    fig, ax = plt.subplots(figsize=(1, 1))
                    ax.text(0.5, 0.5, '测试', fontsize=12)
                    plt.close(fig)
                    print(f"使用中文字体: {font}")
                    return
                except:
                    continue

            # 如果都不行，使用默认字体
            plt.rcParams['font.sans-serif'] = ['DejaVu Sans']
            plt.rcParams['axes.unicode_minus'] = False
            print("警告: 无法加载中文字体，使用默认字体")

        except Exception as e:
            print(f"字体设置警告: {e}")
            plt.rcParams['font.sans-serif'] = ['DejaVu Sans']
            plt.rcParams['axes.unicode_minus'] = False

    def load_data(self):
        """加载数据"""
        try:
            if not os.path.exists(self.data_file):
                raise FileNotFoundError(f"数据文件不存在: {self.data_file}")

            print(f"正在加载数据: {self.data_file}")

            # 根据文件扩展名选择读取方式
            if self.data_file.endswith('.parquet'):
                self.df = pd.read_parquet(self.data_file)
            elif self.data_file.endswith('.csv'):
                self.df = pd.read_csv(self.data_file)
            else:
                raise ValueError("不支持的文件格式，请使用 .parquet 或 .csv 文件")

            print(f"数据加载成功: {len(self.df):,} 条记录")

            # 确保数据按数列号排序
            if '数列号' in self.df.columns:
                self.df = self.df.sort_values('数列号').reset_index(drop=True)
                print("数据已按数列号排序")
            else:
                print("警告: 未找到数列号列，使用原始顺序")

            return True

        except Exception as e:
            print(f"数据加载失败: {e}")
            return False

    def analyze_large_buy1_orders(self, threshold=99):
        """分析买1量大于阈值的数据"""
        print(f"\n=== 分析买1量 > {threshold} 的数据 ===")

        # 找到买1量列
        buy1_col = None
        for col in self.df.columns:
            if '买1量' in col:
                buy1_col = col
                break

        if buy1_col is None:
            raise ValueError("未找到买1量列")

        print(f"使用买1量列: {buy1_col}")

        # 筛选买1量大于阈值的数据
        large_buy1_mask = self.df[buy1_col] > threshold
        self.buy1_large_orders = self.df[large_buy1_mask].copy().reset_index(drop=True)

        print(f"买1量 > {threshold} 的记录数: {len(self.buy1_large_orders):,}")
        print(f"占总记录比例: {len(self.buy1_large_orders)/len(self.df)*100:.2f}%")

        if len(self.buy1_large_orders) == 0:
            print("未找到符合条件的买1量记录")
            return False

        # 显示买1量统计
        print(f"\n买1量统计:")
        print(f"  最小值: {self.buy1_large_orders[buy1_col].min()}")
        print(f"  最大值: {self.buy1_large_orders[buy1_col].max()}")
        print(f"  平均值: {self.buy1_large_orders[buy1_col].mean():.1f}")
        print(f"  中位数: {self.buy1_large_orders[buy1_col].median()}")

        return True

    def analyze_large_sell1_orders(self, threshold=99):
        """分析卖1量大于阈值的数据"""
        print(f"\n=== 分析卖1量 > {threshold} 的数据 ===")

        # 找到卖1量列
        sell1_col = None
        for col in self.df.columns:
            if '卖1量' in col:
                sell1_col = col
                break

        if sell1_col is None:
            raise ValueError("未找到卖1量列")

        print(f"使用卖1量列: {sell1_col}")

        # 筛选卖1量大于阈值的数据
        large_sell1_mask = self.df[sell1_col] > threshold
        self.sell1_large_orders = self.df[large_sell1_mask].copy().reset_index(drop=True)

        print(f"卖1量 > {threshold} 的记录数: {len(self.sell1_large_orders):,}")
        print(f"占总记录比例: {len(self.sell1_large_orders)/len(self.df)*100:.2f}%")

        if len(self.sell1_large_orders) == 0:
            print("未找到符合条件的卖1量记录")
            return False

        # 显示卖1量统计
        print(f"\n卖1量统计:")
        print(f"  最小值: {self.sell1_large_orders[sell1_col].min()}")
        print(f"  最大值: {self.sell1_large_orders[sell1_col].max()}")
        print(f"  平均值: {self.sell1_large_orders[sell1_col].mean():.1f}")
        print(f"  中位数: {self.sell1_large_orders[sell1_col].median()}")

        return True

    def calculate_future_price_ranges(self):
        """计算后续成交价范围 (买1量和卖1量)"""
        print(f"\n=== 计算后续成交价范围 ===")

        # 买1量分析
        if len(self.buy1_large_orders) > 0:
            self._calculate_buy1_ranges()

        # 卖1量分析
        if len(self.sell1_large_orders) > 0:
            self._calculate_sell1_ranges()

        return True

    def _calculate_buy1_ranges(self):
        """计算买1量的后续价格范围"""
        buy1_col = None
        for col in self.df.columns:
            if '买1量' in col:
                buy1_col = col
                break

        buy1_results_100 = []
        buy1_results_200 = []

        for idx, row in self.buy1_large_orders.iterrows():
            current_seq_num = row['数列号'] if '数列号' in row else idx
            current_price = row['成交价']
            buy1_volume = row[buy1_col]

            # 100笔分析
            future_mask_100 = self.df['数列号'] >= current_seq_num + 1
            future_data_100 = self.df[future_mask_100].head(100)

            if len(future_data_100) > 0:
                min_price_100 = future_data_100['成交价'].min()
                max_price_100 = future_data_100['成交价'].max()
                price_range_100 = max_price_100 - min_price_100

                # 正确计算价格变化（相对于当前价格的最低价变化）
                price_decline_100 = current_price - min_price_100  # 下跌幅度（正数表示下跌）
                price_rise_100 = max_price_100 - current_price     # 上涨幅度（正数表示上涨）

                # 计算净变化（最终相对于起始的变化）
                final_price_100 = future_data_100['成交价'].iloc[-1]
                net_change_100 = final_price_100 - current_price
                net_change_pct_100 = (net_change_100 / current_price) * 100

                buy1_results_100.append({
                    '数列号': current_seq_num,
                    '成交价': current_price,
                    '买1量': buy1_volume,
                    '未来100笔最小价': min_price_100,
                    '未来100笔最大价': max_price_100,
                    '未来100笔最终价': final_price_100,
                    '未来100笔价格范围': price_range_100,
                    '未来100笔价格下跌': price_decline_100,
                    '未来100笔价格上涨': price_rise_100,
                    '未来100笔净变化': net_change_100,
                    '未来100笔净变化%': net_change_pct_100,
                    '实际样本数': len(future_data_100)
                })

            # 200笔分析
            future_mask_200 = self.df['数列号'] >= current_seq_num + 1
            future_data_200 = self.df[future_mask_200].head(200)

            if len(future_data_200) > 0:
                min_price_200 = future_data_200['成交价'].min()
                max_price_200 = future_data_200['成交价'].max()
                price_range_200 = max_price_200 - min_price_200

                # 正确计算价格变化
                price_decline_200 = current_price - min_price_200  # 下跌幅度
                price_rise_200 = max_price_200 - current_price     # 上涨幅度

                # 计算净变化
                final_price_200 = future_data_200['成交价'].iloc[-1]
                net_change_200 = final_price_200 - current_price
                net_change_pct_200 = (net_change_200 / current_price) * 100

                buy1_results_200.append({
                    '数列号': current_seq_num,
                    '成交价': current_price,
                    '买1量': buy1_volume,
                    '未来200笔最小价': min_price_200,
                    '未来200笔最大价': max_price_200,
                    '未来200笔最终价': final_price_200,
                    '未来200笔价格范围': price_range_200,
                    '未来200笔价格下跌': price_decline_200,
                    '未来200笔价格上涨': price_rise_200,
                    '未来200笔净变化': net_change_200,
                    '未来200笔净变化%': net_change_pct_200,
                    '实际样本数': len(future_data_200)
                })

        self.buy1_results_100 = pd.DataFrame(buy1_results_100)
        self.buy1_results_200 = pd.DataFrame(buy1_results_200)
        print(f"买1量分析完成: 100笔 {len(self.buy1_results_100)} 条, 200笔 {len(self.buy1_results_200)} 条")

    def _calculate_sell1_ranges(self):
        """计算卖1量的后续价格范围"""
        sell1_col = None
        for col in self.df.columns:
            if '卖1量' in col:
                sell1_col = col
                break

        sell1_results_100 = []
        sell1_results_200 = []

        for idx, row in self.sell1_large_orders.iterrows():
            current_seq_num = row['数列号'] if '数列号' in row else idx
            current_price = row['成交价']
            sell1_volume = row[sell1_col]

            # 100笔分析
            future_mask_100 = self.df['数列号'] >= current_seq_num + 1
            future_data_100 = self.df[future_mask_100].head(100)

            if len(future_data_100) > 0:
                min_price_100 = future_data_100['成交价'].min()
                max_price_100 = future_data_100['成交价'].max()
                price_range_100 = max_price_100 - min_price_100

                # 计算价格变化（相对于当前价格）
                price_decline_100 = current_price - min_price_100  # 下跌幅度
                price_rise_100 = max_price_100 - current_price     # 上涨幅度

                # 计算净变化（最终相对于起始的变化）
                final_price_100 = future_data_100['成交价'].iloc[-1]
                net_change_100 = final_price_100 - current_price
                net_change_pct_100 = (net_change_100 / current_price) * 100

                sell1_results_100.append({
                    '数列号': current_seq_num,
                    '成交价': current_price,
                    '卖1量': sell1_volume,
                    '未来100笔最小价': min_price_100,
                    '未来100笔最大价': max_price_100,
                    '未来100笔最终价': final_price_100,
                    '未来100笔价格范围': price_range_100,
                    '未来100笔价格下跌': price_decline_100,
                    '未来100笔价格上涨': price_rise_100,
                    '未来100笔净变化': net_change_100,
                    '未来100笔净变化%': net_change_pct_100,
                    '实际样本数': len(future_data_100)
                })

            # 200笔分析
            future_mask_200 = self.df['数列号'] >= current_seq_num + 1
            future_data_200 = self.df[future_mask_200].head(200)

            if len(future_data_200) > 0:
                min_price_200 = future_data_200['成交价'].min()
                max_price_200 = future_data_200['成交价'].max()
                price_range_200 = max_price_200 - min_price_200

                # 计算价格变化
                price_decline_200 = current_price - min_price_200  # 下跌幅度
                price_rise_200 = max_price_200 - current_price     # 上涨幅度

                # 计算净变化
                final_price_200 = future_data_200['成交价'].iloc[-1]
                net_change_200 = final_price_200 - current_price
                net_change_pct_200 = (net_change_200 / current_price) * 100

                sell1_results_200.append({
                    '数列号': current_seq_num,
                    '成交价': current_price,
                    '卖1量': sell1_volume,
                    '未来200笔最小价': min_price_200,
                    '未来200笔最大价': max_price_200,
                    '未来200笔最终价': final_price_200,
                    '未来200笔价格范围': price_range_200,
                    '未来200笔价格下跌': price_decline_200,
                    '未来200笔价格上涨': price_rise_200,
                    '未来200笔净变化': net_change_200,
                    '未来200笔净变化%': net_change_pct_200,
                    '实际样本数': len(future_data_200)
                })

        self.sell1_results_100 = pd.DataFrame(sell1_results_100)
        self.sell1_results_200 = pd.DataFrame(sell1_results_200)
        print(f"卖1量分析完成: 100笔 {len(self.sell1_results_100)} 条, 200笔 {len(self.sell1_results_200)} 条")

    def print_analysis_results(self):
        """打印分析结果"""
        print(f"\n" + "="*90)
        print("AU2512期货大单分析结果 (买1量 vs 卖1量)")
        print("="*90)

        # 对比表
        print(f"\n【大单数据对比】")
        print(f"{'类型':>8} {'样本数':>8} {'占比':>8} {'成交量范围':>12} {'平均成交量':>12}")
        print("-" * 60)

        if len(self.buy1_large_orders) > 0:
            buy1_col = None
            for col in self.df.columns:
                if '买1量' in col:
                    buy1_col = col
                    break
            buy_vol_range = f"{self.buy1_large_orders[buy1_col].min()}-{self.buy1_large_orders[buy1_col].max()}"
            buy_avg_vol = f"{self.buy1_large_orders[buy1_col].mean():.1f}"
            print(f"{'买1量':>8} {len(self.buy1_large_orders):>8} {len(self.buy1_large_orders)/len(self.df)*100:>7.2f}% {buy_vol_range:>12} {buy_avg_vol:>12}")

        if len(self.sell1_large_orders) > 0:
            sell1_col = None
            for col in self.df.columns:
                if '卖1量' in col:
                    sell1_col = col
                    break
            sell_vol_range = f"{self.sell1_large_orders[sell1_col].min()}-{self.sell1_large_orders[sell1_col].max()}"
            sell_avg_vol = f"{self.sell1_large_orders[sell1_col].mean():.1f}"
            print(f"{'卖1量':>8} {len(self.sell1_large_orders):>8} {len(self.sell1_large_orders)/len(self.df)*100:>7.2f}% {sell_vol_range:>12} {sell_avg_vol:>12}")

        # 100笔分析对比
        print(f"\n【100笔后续表现对比】")
        print(f"{'类型':>8} {'价格范围':>10} {'趋势变化':>10} {'成功率':>8} {'平均幅度%':>12}")
        print("-" * 60)

        if len(self.buy1_results_100) > 0:
            buy_avg_range = self.buy1_results_100['未来100笔价格范围'].mean()
            buy_avg_net_change = self.buy1_results_100['未来100笔净变化%'].mean()
            buy_decline_rate = (self.buy1_results_100['未来100笔价格下跌'] > 0).sum() / len(self.buy1_results_100) * 100
            trend_buy = "下跌" if buy_avg_net_change < 0 else "上涨"
            print(f"{'买1量':>8} {buy_avg_range:>10.2f} {trend_buy:>10} {buy_decline_rate:>7.1f}% {buy_avg_net_change:+>11.3f}%")

        if len(self.sell1_results_100) > 0:
            sell_avg_range = self.sell1_results_100['未来100笔价格范围'].mean()
            sell_avg_net_change = self.sell1_results_100['未来100笔净变化%'].mean()
            sell_decline_rate = (self.sell1_results_100['未来100笔价格下跌'] > 0).sum() / len(self.sell1_results_100) * 100
            trend_sell = "下跌" if sell_avg_net_change < 0 else "上涨"
            print(f"{'卖1量':>8} {sell_avg_range:>10.2f} {trend_sell:>10} {sell_decline_rate:>7.1f}% {sell_avg_net_change:+>11.3f}%")

        # 200笔分析对比
        print(f"\n【200笔后续表现对比】")
        print(f"{'类型':>8} {'价格范围':>10} {'趋势变化':>10} {'成功率':>8} {'平均幅度%':>12}")
        print("-" * 60)

        if len(self.buy1_results_200) > 0:
            buy_avg_range_200 = self.buy1_results_200['未来200笔价格范围'].mean()
            buy_avg_change_200 = self.buy1_results_200['未来200笔净变化%'].mean()
            buy_decline_rate_200 = (self.buy1_results_200['未来200笔价格下跌'] > 0).sum() / len(self.buy1_results_200) * 100
            trend_buy_200 = "下跌" if buy_avg_change_200 < 0 else "上涨"
            print(f"{'买1量':>8} {buy_avg_range_200:>10.2f} {trend_buy_200:>10} {buy_decline_rate_200:>7.1f}% {buy_avg_change_200:+>11.3f}%")

        if len(self.sell1_results_200) > 0:
            sell_avg_range_200 = self.sell1_results_200['未来200笔价格范围'].mean()
            sell_avg_change_200 = self.sell1_results_200['未来200笔净变化%'].mean()
            sell_decline_rate_200 = (self.sell1_results_200['未来200笔价格下跌'] > 0).sum() / len(self.sell1_results_200) * 100
            trend_sell_200 = "下跌" if sell_avg_change_200 < 0 else "上涨"
            print(f"{'卖1量':>8} {sell_avg_range_200:>10.2f} {trend_sell_200:>10} {sell_decline_rate_200:>7.1f}% {sell_avg_change_200:+>11.3f}%")

        # 详细样本展示
        print(f"\n【详细分析 - 各类型前5个样本】")
        print("\n买1量大单样本:")
        if len(self.buy1_results_100) > 0:
            print(f"{'序号':>4} {'数列号':>8} {'成交价':>8} {'买1量':>8} {'100笔范围':>10} {'100笔涨幅%':>12}")
            print("-" * 60)
            for i in range(min(5, len(self.buy1_results_100))):
                row = self.buy1_results_100.iloc[i]
                print(f"{i+1:>4} {row['数列号']:>8.0f} {row['成交价']:>8.2f} {row['买1量']:>8.0f} "
                      f"{row['未来100笔价格范围']:>10.2f} {row['未来100笔净变化%']:+>12.3f}%")

        print("\n卖1量大单样本:")
        if len(self.sell1_results_100) > 0:
            print(f"{'序号':>4} {'数列号':>8} {'成交价':>8} {'卖1量':>8} {'100笔范围':>10} {'100笔变化%':>12}")
            print("-" * 60)
            for i in range(min(5, len(self.sell1_results_100))):
                row = self.sell1_results_100.iloc[i]
                print(f"{i+1:>4} {row['数列号']:>8.0f} {row['成交价']:>8.2f} {row['卖1量']:>8.0f} "
                      f"{row['未来100笔价格范围']:>10.2f} {row['未来100笔净变化%']:+>12.3f}%")

        # 业务洞察
        print(f"\n【业务洞察】")
        print(f"1. 大单样本数量对比:")
        if len(self.buy1_large_orders) > 0 and len(self.sell1_large_orders) > 0:
            print(f"   - 买1量大单: {len(self.buy1_large_orders)} 个 ({len(self.buy1_large_orders)/len(self.df)*100:.2f}%)")
            print(f"   - 卖1量大单: {len(self.sell1_large_orders)} 个 ({len(self.sell1_large_orders)/len(self.df)*100:.2f}%)")
            ratio = len(self.buy1_large_orders) / len(self.sell1_large_orders)
            print(f"   - 买卖单比例: {ratio:.1f}:1 (买单占{'主导' if ratio > 1 else '劣势'})")

        print(f"2. 后续价格波动分析:")
        if len(self.buy1_results_100) > 0 and len(self.sell1_results_100) > 0:
            buy_volatility = self.buy1_results_100['未来100笔价格范围'].mean()
            sell_volatility = self.sell1_results_100['未来100笔价格范围'].mean()
            print(f"   - 买1量大单后平均波动: {buy_volatility:.2f}元")
            print(f"   - 卖1量大单后平均波动: {sell_volatility:.2f}元")
            higher_volatility = "买1量" if buy_volatility > sell_volatility else "卖1量"
            print(f"   - {higher_volatility}大单引发更大的价格波动")

        print(f"3. 趋势预测能力:")
        # 判断主要趋势方向
        overall_trend = "强势上涨" if self._is_overall_uptrend() else "震荡整理"
        print(f"   - 市场整体趋势: {overall_trend}")
        print(f"   - 大单信号强度: {'极强' if self._is_signal_strong() else '中等' if self._is_signal_moderate() else '较弱'}")

    def _is_overall_uptrend(self):
        """判断是否为整体上涨趋势"""
        buy_positive = (self.buy1_results_100['未来100笔净变化'] > 0).sum() / len(self.buy1_results_100) if len(self.buy1_results_100) > 0 else 0
        sell_positive = (self.sell1_results_100['未来100笔净变化'] > 0).sum() / len(self.sell1_results_100) if len(self.sell1_results_100) > 0 else 0

        # 如果买1量大单和卖1量大单后都主要上涨，说明整体强势
        return buy_positive > 0.5 and sell_positive > 0.5

    def _is_signal_strong(self):
        """判断信号强度"""
        if len(self.buy1_results_100) > 0:
            buy_decline_rate = (self.buy1_results_100['未来100笔价格下跌'] > 0).sum() / len(self.buy1_results_100)
            # 如果下跌概率很高，说明信号强度强
            return buy_decline_rate > 0.8
        return False

    def _is_signal_moderate(self):
        """判断信号强度为中等"""
        if len(self.buy1_results_100) > 0:
            buy_decline_rate = (self.buy1_results_100['未来100笔价格下跌'] > 0).sum() / len(self.buy1_results_100)
            return 0.5 < buy_decline_rate <= 0.8
        return False

    def create_comprehensive_chart(self):
        """创建综合对比图表"""
        print(f"\n=== 生成综合分析图表 ===")

        try:
            # 创建大型综合图表
            fig = plt.figure(figsize=(24, 16))
            gs = fig.add_gridspec(4, 4, hspace=0.3, wspace=0.3)

            # 1. 样本数量对比 (左上)
            ax1 = fig.add_subplot(gs[0, 0])
            order_types = []
            sample_counts = []
            colors = []

            if len(self.buy1_large_orders) > 0:
                order_types.append('买1量大单')
                sample_counts.append(len(self.buy1_large_orders))
                colors.append('lightgreen')
            if len(self.sell1_large_orders) > 0:
                order_types.append('卖1量大单')
                sample_counts.append(len(self.sell1_large_orders))
                colors.append('lightcoral')

            bars = ax1.bar(order_types, sample_counts, color=colors, alpha=0.8, edgecolor='black')
            ax1.set_ylabel('样本数量', fontsize=12)
            ax1.set_title('大单样本数量对比', fontsize=14, fontweight='bold')
            ax1.grid(True, alpha=0.3, axis='y')

            # 添加数值标签
            for bar, count in zip(bars, sample_counts):
                height = bar.get_height()
                ax1.text(bar.get_x() + bar.get_width()/2., height + 1,
                        f'{count}', ha='center', va='bottom', fontsize=12, fontweight='bold')

            # 2. 价格范围对比 (右上)
            ax2 = fig.add_subplot(gs[0, 1])
            categories = []
            buy_ranges = []
            sell_ranges = []

            if len(self.buy1_results_100) > 0:
                categories.append('100笔')
                buy_ranges.append(self.buy1_results_100['未来100笔价格范围'].mean())
            if len(self.buy1_results_200) > 0:
                categories.append('200笔')
                buy_ranges.append(self.buy1_results_200['未来200笔价格范围'].mean())

            x = np.arange(len(categories))
            width = 0.35

            if len(self.buy1_results_100) > 0:
                ax2.bar(x - width/2, buy_ranges, width, label='买1量', color='lightgreen', alpha=0.8)

            if len(self.sell1_results_100) > 0:
                sell_ranges = []
                if len(self.sell1_results_100) > 0:
                    sell_ranges.append(self.sell1_results_100['未来100笔价格范围'].mean())
                if len(self.sell1_results_200) > 0:
                    sell_ranges.append(self.sell1_results_200['未来200笔价格范围'].mean())
                ax2.bar(x + width/2, sell_ranges, width, label='卖1量', color='lightcoral', alpha=0.8)

            ax2.set_xlabel('时间窗口', fontsize=12)
            ax2.set_ylabel('平均价格范围 (元)', fontsize=12)
            ax2.set_title('价格波动范围对比', fontsize=14, fontweight='bold')
            ax2.set_xticks(x)
            ax2.set_xticklabels(categories)
            ax2.legend()
            ax2.grid(True, alpha=0.3, axis='y')

            # 3. 成功率对比 (左下第二个)
            ax3 = fig.add_subplot(gs[0, 2])
            success_types = []
            success_rates = []
            success_colors = []

            if len(self.buy1_results_100) > 0:
                buy_decline_rate = (self.buy1_results_100['未来100笔价格下跌'] > 0).sum() / len(self.buy1_results_100) * 100
                success_types.append('买1量\n下跌概率')
                success_rates.append(buy_decline_rate)
                success_colors.append('lightgreen')

            if len(self.sell1_results_100) > 0:
                sell_decline_rate = (self.sell1_results_100['未来100笔价格下跌'] > 0).sum() / len(self.sell1_results_100) * 100
                success_types.append('卖1量\n下跌概率')
                success_rates.append(sell_decline_rate)
                success_colors.append('lightcoral')

            bars = ax3.bar(success_types, success_rates, color=success_colors, alpha=0.8, edgecolor='black')
            ax3.set_ylabel('下跌概率 (%)', fontsize=12)
            ax3.set_title('价格下跌概率对比', fontsize=14, fontweight='bold')
            ax3.set_ylim(0, 105)
            ax3.grid(True, alpha=0.3, axis='y')

            # 添加数值标签
            for bar, rate in zip(bars, success_rates):
                height = bar.get_height()
                ax3.text(bar.get_x() + bar.get_width()/2., height - 3,
                        f'{rate:.1f}%', ha='center', va='top', fontsize=11, fontweight='bold')

            # 4. 散点图对比 (右下第二个)
            ax4 = fig.add_subplot(gs[0, 3])
            if len(self.buy1_results_100) > 0:
                ax4.scatter(self.buy1_results_100['买1量'], self.buy1_results_100['未来100笔价格范围'],
                          alpha=0.7, color='green', label='买1量', s=50)
            if len(self.sell1_results_100) > 0:
                ax4.scatter(self.sell1_results_100['卖1量'], self.sell1_results_100['未来100笔价格范围'],
                          alpha=0.7, color='red', label='卖1量', s=50)
            ax4.set_xlabel('挂单量 (手)', fontsize=12)
            ax4.set_ylabel('价格范围 (元)', fontsize=12)
            ax4.set_title('挂单量 vs 价格波动', fontsize=14, fontweight='bold')
            ax4.legend()
            ax4.grid(True, alpha=0.3)

            # 5-6. 100笔大单价格变化幅度柱状图 (第二行)
            # 买1量100笔价格变化幅度 (第二行左半部分)
            if len(self.buy1_results_100) > 0:
                ax5 = fig.add_subplot(gs[1, 0])
                sample_count = min(25, len(self.buy1_results_100))  # 显示前25个样本
                sample_indices = range(1, sample_count + 1)

                # 买1量100笔价格净变化（元）
                buy1_changes_100 = self.buy1_results_100['未来100笔净变化'].iloc[:sample_count].values
                # 为柱子分配颜色：正值为绿色，负值为红色
                buy1_colors_100 = ['lightgreen' if x >= 0 else 'lightcoral' for x in buy1_changes_100]

                buy1_bars_100 = ax5.bar(sample_indices, buy1_changes_100,
                                       width=0.6, alpha=0.8, color=buy1_colors_100,
                                       label='买1量100笔价格变化(元)', edgecolor='black', linewidth=0.5)

                # 添加零线
                ax5.axhline(y=0, color='black', linestyle='-', linewidth=1, alpha=0.7)

                ax5.set_xlabel('样本序号', fontsize=12)
                ax5.set_ylabel('价格变化幅度 (元)', fontsize=12)
                ax5.set_title('买1量大单 - 100笔价格变化', fontsize=14, fontweight='bold')
                ax5.grid(True, alpha=0.3, axis='y')

                # 添加数值标签
                for i, change in enumerate(buy1_changes_100):
                    y_pos = change + 0.01 if change >= 0 else change - 0.01
                    ax5.text(i + 1, y_pos, f'{change:+.2f}',
                           ha='center', va='bottom' if change >= 0 else 'top',
                           fontsize=9, rotation=45)

            # 卖1量100笔价格变化幅度 (第二行右半部分)
            if len(self.sell1_results_100) > 0:
                ax6 = fig.add_subplot(gs[1, 1])
                sample_count = min(25, len(self.sell1_results_100))  # 显示前25个样本
                sample_indices = range(1, sample_count + 1)

                # 卖1量100笔价格净变化（元）
                sell1_changes_100 = self.sell1_results_100['未来100笔净变化'].iloc[:sample_count].values
                # 为柱子分配颜色：正值为绿色，负值为红色
                sell1_colors_100 = ['lightgreen' if x >= 0 else 'lightcoral' for x in sell1_changes_100]

                sell1_bars_100 = ax6.bar(sample_indices, sell1_changes_100,
                                        width=0.6, alpha=0.8, color=sell1_colors_100,
                                        label='卖1量100笔价格变化(元)', edgecolor='black', linewidth=0.5)

                # 添加零线
                ax6.axhline(y=0, color='black', linestyle='-', linewidth=1, alpha=0.7)

                ax6.set_xlabel('样本序号', fontsize=12)
                ax6.set_ylabel('价格变化幅度 (元)', fontsize=12)
                ax6.set_title('卖1量大单 - 100笔价格变化', fontsize=14, fontweight='bold')
                ax6.grid(True, alpha=0.3, axis='y')

                # 添加数值标签
                for i, change in enumerate(sell1_changes_100):
                    y_pos = change + 0.01 if change >= 0 else change - 0.01
                    ax6.text(i + 1, y_pos, f'{change:+.2f}',
                           ha='center', va='bottom' if change >= 0 else 'top',
                           fontsize=9, rotation=45)

            # 7-8. 200笔大单价格变化幅度柱状图 (第二行第三、四列)
            # 买1量200笔价格变化幅度 (第二行第三个)
            if len(self.buy1_results_200) > 0:
                ax7 = fig.add_subplot(gs[1, 2])
                sample_count = min(25, len(self.buy1_results_200))  # 显示前25个样本
                sample_indices = range(1, sample_count + 1)

                # 买1量200笔价格净变化（元）
                buy1_changes_200 = self.buy1_results_200['未来200笔净变化'].iloc[:sample_count].values
                # 为柱子分配颜色：正值为绿色，负值为红色
                buy1_colors_200 = ['darkgreen' if x >= 0 else 'darkred' for x in buy1_changes_200]

                buy1_bars_200 = ax7.bar(sample_indices, buy1_changes_200,
                                       width=0.6, alpha=0.8, color=buy1_colors_200,
                                       label='买1量200笔价格变化(元)', edgecolor='black', linewidth=0.5)

                # 添加零线
                ax7.axhline(y=0, color='black', linestyle='-', linewidth=1, alpha=0.7)

                ax7.set_xlabel('样本序号', fontsize=12)
                ax7.set_ylabel('价格变化幅度 (元)', fontsize=12)
                ax7.set_title('买1量大单 - 200笔价格变化', fontsize=14, fontweight='bold')
                ax7.grid(True, alpha=0.3, axis='y')

                # 添加数值标签
                for i, change in enumerate(buy1_changes_200):
                    y_pos = change + 0.01 if change >= 0 else change - 0.01
                    ax7.text(i + 1, y_pos, f'{change:+.2f}',
                           ha='center', va='bottom' if change >= 0 else 'top',
                           fontsize=9, rotation=45)

            # 卖1量200笔价格变化幅度 (第二行第四个)
            if len(self.sell1_results_200) > 0:
                ax8 = fig.add_subplot(gs[1, 3])
                sample_count = min(25, len(self.sell1_results_200))  # 显示前25个样本
                sample_indices = range(1, sample_count + 1)

                # 卖1量200笔价格净变化（元）
                sell1_changes_200 = self.sell1_results_200['未来200笔净变化'].iloc[:sample_count].values
                # 为柱子分配颜色：正值为绿色，负值为红色
                sell1_colors_200 = ['darkgreen' if x >= 0 else 'darkred' for x in sell1_changes_200]

                sell1_bars_200 = ax8.bar(sample_indices, sell1_changes_200,
                                        width=0.6, alpha=0.8, color=sell1_colors_200,
                                        label='卖1量200笔价格变化(元)', edgecolor='black', linewidth=0.5)

                # 添加零线
                ax8.axhline(y=0, color='black', linestyle='-', linewidth=1, alpha=0.7)

                ax8.set_xlabel('样本序号', fontsize=12)
                ax8.set_ylabel('价格变化幅度 (元)', fontsize=12)
                ax8.set_title('卖1量大单 - 200笔价格变化', fontsize=14, fontweight='bold')
                ax8.grid(True, alpha=0.3, axis='y')

                # 添加数值标签
                for i, change in enumerate(sell1_changes_200):
                    y_pos = change + 0.01 if change >= 0 else change - 0.01
                    ax8.text(i + 1, y_pos, f'{change:+.2f}',
                           ha='center', va='bottom' if change >= 0 else 'top',
                           fontsize=9, rotation=45)

            # 9-12. 箱线图对比 (第三行)
            # 买1量箱线图 (第三行左半部分)
            if len(self.buy1_results_100) > 0:
                ax7 = fig.add_subplot(gs[2, :2])
                buy_data_to_plot = []
                buy_labels = []

                if len(self.buy1_results_100) > 0:
                    buy_data_to_plot.append(self.buy1_results_100['未来100笔价格范围'])
                    buy_labels.append('100笔价格范围')
                if len(self.buy1_results_200) > 0:
                    buy_data_to_plot.append(self.buy1_results_200['未来200笔价格范围'])
                    buy_labels.append('200笔价格范围')
                if len(self.buy1_results_100) > 0:
                    buy_data_to_plot.append(self.buy1_results_100['未来100笔净变化%'] * 100)
                    buy_labels.append('100笔涨幅%(×10)')
                if len(self.buy1_results_200) > 0:
                    buy_data_to_plot.append(self.buy1_results_200['未来200笔净变化%'] * 100)
                    buy_labels.append('200笔涨幅%(×10)')

                bp = ax7.boxplot(buy_data_to_plot, labels=buy_labels, patch_artist=True)
                colors = ['lightgreen', 'darkgreen', 'lightblue', 'blue']
                for patch, color in zip(bp['boxes'], colors[:len(bp['boxes'])]):
                    patch.set_facecolor(color)

                ax7.set_ylabel('数值', fontsize=12)
                ax7.set_title('买1量大单 - 分布特征', fontsize=14, fontweight='bold')
                ax7.grid(True, alpha=0.3)

            # 卖1量箱线图 (第三行右半部分)
            if len(self.sell1_results_100) > 0:
                ax8 = fig.add_subplot(gs[2, 2:])
                sell_data_to_plot = []
                sell_labels = []

                if len(self.sell1_results_100) > 0:
                    sell_data_to_plot.append(self.sell1_results_100['未来100笔价格范围'])
                    sell_labels.append('100笔价格范围')
                if len(self.sell1_results_200) > 0:
                    sell_data_to_plot.append(self.sell1_results_200['未来200笔价格范围'])
                    sell_labels.append('200笔价格范围')
                if len(self.sell1_results_100) > 0:
                    sell_data_to_plot.append(self.sell1_results_100['未来100笔净变化%'] * 100)
                    sell_labels.append('100笔跌幅%(×10)')
                if len(self.sell1_results_200) > 0:
                    sell_data_to_plot.append(self.sell1_results_200['未来200笔净变化%'] * 100)
                    sell_labels.append('200笔跌幅%(×10)')

                bp = ax8.boxplot(sell_data_to_plot, labels=sell_labels, patch_artist=True)
                colors = ['lightcoral', 'darkred', 'lightpink', 'red']
                for patch, color in zip(bp['boxes'], colors[:len(bp['boxes'])]):
                    patch.set_facecolor(color)

                ax8.set_ylabel('数值', fontsize=12)
                ax8.set_title('卖1量大单 - 分布特征', fontsize=14, fontweight='bold')
                ax8.grid(True, alpha=0.3)

            # 13. 综合摘要 (第四行)
            ax9 = fig.add_subplot(gs[3, :])
            ax9.axis('off')

            # 创建综合摘要文本
            summary_text = self._generate_summary_text()

            # 显示摘要文本
            ax9.text(0.05, 0.95, summary_text, transform=ax9.transAxes,
                    fontsize=10, ha='left', va='top',
                    bbox=dict(boxstyle='round,pad=0.5', facecolor='lightyellow', alpha=0.3))

            # 设置总标题
            fig.suptitle('AU2512期货大单综合分析报告 (买1量 vs 卖1量)', fontsize=22, fontweight='bold', y=0.98)

            # 调整布局
            plt.tight_layout()

            # 保存图表
            chart_file = os.path.join(self.output_dir, 'large_orders_comprehensive_analysis.png')
            plt.savefig(chart_file, dpi=300, bbox_inches='tight', facecolor='white', edgecolor='none')
            plt.close()

            print(f"综合分析图表已保存: {chart_file}")

            return True

        except Exception as e:
            print(f"创建综合图表时发生错误: {e}")
            return False

    def _generate_summary_text(self):
        """生成综合摘要文本"""
        text_parts = ["大单综合分析摘要\n"]

        # 基础统计
        text_parts.append("=" * 50)
        text_parts.append("样本统计:")

        if len(self.buy1_large_orders) > 0:
            buy1_col = None
            for col in self.df.columns:
                if '买1量' in col:
                    buy1_col = col
                    break
            text_parts.append(f"• 买1量大单: {len(self.buy1_large_orders)} 个 ({len(self.buy1_large_orders)/len(self.df)*100:.2f}%)")
            text_parts.append(f"  买1量范围: {self.buy1_large_orders[buy1_col].min()}-{self.buy1_large_orders[buy1_col].max()} 手")

        if len(self.sell1_large_orders) > 0:
            sell1_col = None
            for col in self.df.columns:
                if '卖1量' in col:
                    sell1_col = col
                    break
            text_parts.append(f"• 卖1量大单: {len(self.sell1_large_orders)} 个 ({len(self.sell1_large_orders)/len(self.df)*100:.2f}%)")
            text_parts.append(f"  卖1量范围: {self.sell1_large_orders[sell1_col].min()}-{self.sell1_large_orders[sell1_col].max()} 手")

        # 表现对比
        text_parts.append("\n" + "=" * 50)
        text_parts.append("表现对比:")

        if len(self.buy1_results_100) > 0:
            buy_avg_range = self.buy1_results_100['未来100笔价格范围'].mean()
            buy_avg_change = self.buy1_results_100['未来100笔最大涨幅%'].mean()
            buy_success_rate = (self.buy1_results_100['未来100笔最大涨幅'] > 0).sum() / len(self.buy1_results_100) * 100
            text_parts.append(f"• 买1量大单后:")
            text_parts.append(f"  - 平均价格波动: {buy_avg_range:.3f} 元")
            text_parts.append(f"  - 平均涨幅: {buy_avg_change:+.3f}%")
            text_parts.append(f"  - 上涨概率: {buy_success_rate:.1f}%")

        if len(self.sell1_results_100) > 0:
            sell_avg_range = self.sell1_results_100['未来100笔价格范围'].mean()
            sell_avg_change = self.sell1_results_100['未来100笔价格变化%'].mean()
            sell_success_rate = (self.sell1_results_100['未来100笔价格变化'] < 0).sum() / len(self.sell1_results_100) * 100
            text_parts.append(f"• 卖1量大单后:")
            text_parts.append(f"  - 平均价格波动: {sell_avg_range:.3f} 元")
            text_parts.append(f"  - 平均变化: {sell_avg_change:+.3f}%")
            text_parts.append(f"  - 下跌概率: {sell_success_rate:.1f}%")

        # 市场洞察
        text_parts.append("\n" + "=" * 50)
        text_parts.append("市场洞察:")

        # 判断市场特征
        if len(self.buy1_results_100) > 0 and len(self.sell1_results_100) > 0:
            buy_success = (self.buy1_results_100['未来100笔最大涨幅'] > 0).sum() / len(self.buy1_results_100)
            sell_decline = (self.sell1_results_100['未来100笔价格变化'] < 0).sum() / len(self.sell1_results_100)

            if buy_success > 0.9 and sell_decline < 0.3:
                text_parts.append("• 市场特征: 强势上涨市场，买盘力量占主导")
            elif buy_success > 0.7 and sell_decline > 0.7:
                text_parts.append("• 市场特征: 震荡市场，多空力量均衡")
            else:
                text_parts.append("• 市场特征: 趋势不明，需进一步观察")

        # 交易建议
        text_parts.append("\n" + "=" * 50)
        text_parts.append("交易建议:")
        text_parts.append("• 大单信号具有较高的预测价值")
        text_parts.append("• 买1量大单可作为价格上涨的参考信号")
        text_parts.append("• 在强势市场中，卖1量大单可能被完全吸收")
        text_parts.append("• 建议结合其他技术指标综合判断")

        return "\n".join(text_parts)

    def save_results(self):
        """保存分析结果"""
        try:
            # 保存买1量详细结果
            if len(self.buy1_results_100) > 0:
                self.buy1_results_100.to_csv(os.path.join(self.output_dir, 'buy1_large_orders_100_analysis.csv'),
                                            index=False, encoding='utf-8-sig')
                self.buy1_results_200.to_csv(os.path.join(self.output_dir, 'buy1_large_orders_200_analysis.csv'),
                                            index=False, encoding='utf-8-sig')

            # 保存卖1量详细结果
            if len(self.sell1_results_100) > 0:
                self.sell1_results_100.to_csv(os.path.join(self.output_dir, 'sell1_large_orders_100_analysis.csv'),
                                             index=False, encoding='utf-8-sig')
                self.sell1_results_200.to_csv(os.path.join(self.output_dir, 'sell1_large_orders_200_analysis.csv'),
                                             index=False, encoding='utf-8-sig')

            # 保存汇总结果
            summary_data = {
                '指标': [],
                '数值': []
            }

            if len(self.buy1_large_orders) > 0:
                summary_data['指标'].extend([
                    '买1量>99的样本数',
                    '买1量100笔平均价格范围',
                    '买1量200笔平均价格范围',
                    '买1量100笔平均净变化%',
                    '买1量100笔下跌概率%'
                ])
                summary_data['数值'].extend([
                    len(self.buy1_large_orders),
                    f"{self.buy1_results_100['未来100笔价格范围'].mean():.2f}",
                    f"{self.buy1_results_200['未来200笔价格范围'].mean():.2f}",
                    f"{self.buy1_results_100['未来100笔净变化%'].mean():+.2f}%",
                    f"{(self.buy1_results_100['未来100笔价格下跌'] > 0).sum()/len(self.buy1_results_100)*100:.1f}%"
                ])

            if len(self.sell1_large_orders) > 0:
                summary_data['指标'].extend([
                    '卖1量>99的样本数',
                    '卖1量100笔平均价格范围',
                    '卖1量200笔平均价格范围',
                    '卖1量100笔平均净变化%',
                    '卖1量100笔下跌概率%'
                ])
                summary_data['数值'].extend([
                    len(self.sell1_large_orders),
                    f"{self.sell1_results_100['未来100笔价格范围'].mean():.2f}",
                    f"{self.sell1_results_200['未来200笔价格范围'].mean():.2f}",
                    f"{self.sell1_results_100['未来100笔净变化%'].mean():+.2f}%",
                    f"{(self.sell1_results_100['未来100笔价格下跌'] > 0).sum()/len(self.sell1_results_100)*100:.1f}%"
                ])

            summary_df = pd.DataFrame(summary_data)
            summary_df.to_csv(os.path.join(self.output_dir, 'large_orders_summary.csv'),
                             index=False, encoding='utf-8-sig')

            print(f"\n分析结果已保存到 {self.output_dir} 目录:")
            if len(self.buy1_results_100) > 0:
                print(f"  buy1_large_orders_100_analysis.csv - 买1量100笔详细分析")
                print(f"  buy1_large_orders_200_analysis.csv - 买1量200笔详细分析")
            if len(self.sell1_results_100) > 0:
                print(f"  sell1_large_orders_100_analysis.csv - 卖1量100笔详细分析")
                print(f"  sell1_large_orders_200_analysis.csv - 卖1量200笔详细分析")
            print(f"  large_orders_summary.csv - 综合汇总分析")

            # 创建综合可视化图表
            self.create_comprehensive_chart()

        except Exception as e:
            print(f"保存结果时发生错误: {e}")

    def run_analysis(self, threshold=99):
        """运行完整分析"""
        print(f"开始AU2512期货大单综合分析 (阈值: {threshold})...")

        # 加载数据
        if not self.load_data():
            return False

        # 分析买1量大单
        buy1_success = self.analyze_large_buy1_orders(threshold)

        # 分析卖1量大单
        sell1_success = self.analyze_large_sell1_orders(threshold)

        if not buy1_success and not sell1_success:
            print("未找到任何符合条件的记录")
            return False

        # 计算后续价格范围
        self.calculate_future_price_ranges()

        # 打印分析结果
        self.print_analysis_results()

        # 保存结果
        self.save_results()

        return True

def main():
    """主函数"""
    parser = argparse.ArgumentParser(description='AU2512期货大单综合分析工具')
    parser.add_argument('data_file', nargs='?',
                       default='data/au2512_20251013.parquet',
                       help='数据文件路径 (默认: data/au2512_20251013.parquet)')
    parser.add_argument('--threshold', '-t', type=int, default=99,
                       help='买卖单量阈值 (默认: 99)')
    parser.add_argument('--output-dir', '-o', default='large_orders',
                       help='输出目录 (默认: large_orders)')

    args = parser.parse_args()

    # 检查数据文件是否存在
    if not os.path.exists(args.data_file):
        print(f"错误: 数据文件不存在 - {args.data_file}")
        print("\n使用方法:")
        print("  python analyze_large_orders.py [数据文件路径] [--threshold 阈值] [--output-dir 输出目录]")
        print(f"  python analyze_large_orders.py  # 使用默认参数")
        sys.exit(1)

    # 创建分析器并运行
    analyzer = LargeOrdersAnalyzer(args.data_file, args.output_dir)
    success = analyzer.run_analysis(args.threshold)

    if success:
        print(f"\n分析完成！结果已保存到: {args.output_dir}")
    else:
        print(f"\n分析失败！")
        sys.exit(1)

if __name__ == "__main__":
    main()