#!/usr/bin/env python3
"""
精细化分析：寻找最符合27-28岁统计的皇帝群体
"""

import sys
import os
import statistics
sys.path.append(os.path.dirname(os.path.abspath(__file__)))

from data.emperors.northern_wei_emperors import NORTHERN_WEI_EMPERORS
from analysis.models import ReliabilityLevel

def refined_analysis():
    """精细化分析不同皇帝群体"""
    
    print("=" * 70)
    print("🔍 精细化分析：寻找27-28岁统计的准确样本")
    print("=" * 70)
    print()
    
    # 所有孝文帝改革前的皇帝
    pre_reform = [emp for emp in NORTHERN_WEI_EMPERORS 
                  if emp.name != "孝文帝拓跋宏" and emp.lifespan is not None]
    
    # 1. 只看高可靠性史料
    high_reliability = [emp for emp in pre_reform if emp.reliability == ReliabilityLevel.HIGH]
    
    print("📊 方案1: 仅高可靠性史料 (★)")
    print("-" * 50)
    if high_reliability:
        lifespans = [emp.lifespan for emp in high_reliability]
        sample_names = [f'{emp.name.split("拓跋")[0]}({emp.lifespan})' for emp in high_reliability]
        print(f"样本: {sample_names}")
        print(f"平均: {statistics.mean(lifespans):.1f}岁, 中位: {statistics.median(lifespans):.1f}岁")
    print()
    
    # 2. 排除太武帝（44岁异常高）
    high_without_taiwu = [emp for emp in high_reliability if "太武帝" not in emp.name]
    
    print("📊 方案2: 高可靠性史料，排除太武帝异常值")
    print("-" * 50)
    if high_without_taiwu:
        lifespans = [emp.lifespan for emp in high_without_taiwu]
        sample_names = [f'{emp.name.split("拓跋")[0]}({emp.lifespan})' for emp in high_without_taiwu]
        print(f"样本: {sample_names}")
        print(f"平均: {statistics.mean(lifespans):.1f}岁, 中位: {statistics.median(lifespans):.1f}岁")
    print()
    
    # 3. 只看短命皇帝（可能更能反映"基因焦虑"）
    short_lived_reliable = [emp for emp in high_reliability if emp.lifespan < 35]
    
    print("📊 方案3: 高可靠性史料中的短命皇帝 (<35岁)")
    print("-" * 50)
    if short_lived_reliable:
        lifespans = [emp.lifespan for emp in short_lived_reliable]
        sample_names = [f'{emp.name.split("拓跋")[0]}({emp.lifespan})' for emp in short_lived_reliable]
        print(f"样本: {sample_names}")
        print(f"平均: {statistics.mean(lifespans):.1f}岁, 中位: {statistics.median(lifespans):.1f}岁")
    print()
    
    # 4. 中后期皇帝（可能更能反映"基因焦虑"加剧）
    later_emperors = [emp for emp in high_reliability 
                     if any(year in emp.reign_period for year in ["452", "465", "471"])]
    
    print("📊 方案4: 中后期高可靠性皇帝 (452年后)")
    print("-" * 50)
    if later_emperors:
        lifespans = [emp.lifespan for emp in later_emperors]
        sample_names = [f'{emp.name.split("拓跋")[0]}({emp.lifespan})' for emp in later_emperors]
        print(f"样本: {sample_names}")
        print(f"平均: {statistics.mean(lifespans):.1f}岁, 中位: {statistics.median(lifespans):.1f}岁")
    print()
    
    # 5. 包含中等可靠性，但排除异常值
    medium_high_reasonable = [emp for emp in pre_reform 
                             if emp.reliability in [ReliabilityLevel.HIGH, ReliabilityLevel.MEDIUM]
                             and emp.lifespan < 45]  # 排除44岁以上的
    
    print("📊 方案5: 中高可靠性，排除异常长寿 (<45岁)")
    print("-" * 50)
    if medium_high_reasonable:
        lifespans = [emp.lifespan for emp in medium_high_reasonable]
        print(f"样本数: {len(medium_high_reasonable)}位")
        for emp in medium_high_reasonable:
            mark = "★" if emp.reliability == ReliabilityLevel.HIGH else "☆"
            print(f"  {mark} {emp.name.split('拓跋')[0]}: {emp.lifespan}岁")
        print(f"平均: {statistics.mean(lifespans):.1f}岁, 中位: {statistics.median(lifespans):.1f}岁")
    print()
    
    # 6. 尝试不同的统计方法
    print("📊 方案6: 不同统计方法对比")
    print("-" * 50)
    if high_reliability:
        lifespans = [emp.lifespan for emp in high_reliability]
        
        # 去掉最高值和最低值
        if len(lifespans) >= 3:
            trimmed = sorted(lifespans)[1:-1]
            print(f"去极值后平均: {statistics.mean(trimmed):.1f}岁")
        
        # 加权平均（按史料可靠性）
        weighted_sum = 0
        weight_sum = 0
        for emp in high_reliability:
            weight = 1.0  # 高可靠性权重为1
            weighted_sum += emp.lifespan * weight
            weight_sum += weight
        
        print(f"加权平均: {weighted_sum/weight_sum:.1f}岁")
        
        # 众数分析
        lifespan_ranges = []
        for lifespan in lifespans:
            if lifespan < 25:
                lifespan_ranges.append("20-25")
            elif lifespan < 30:
                lifespan_ranges.append("25-30")
            elif lifespan < 35:
                lifespan_ranges.append("30-35")
            else:
                lifespan_ranges.append("35+")
        
        from collections import Counter
        range_counts = Counter(lifespan_ranges)
        print(f"年龄段分布: {dict(range_counts)}")
    print()
    
    # 7. 寻找最接近27-28岁的组合
    print("🎯 寻找最接近27-28岁的组合:")
    print("-" * 50)
    
    target_range = (27, 28)
    
    # 尝试不同组合
    combinations = [
        ("仅短命高可靠性", [emp for emp in high_reliability if emp.lifespan < 30]),
        ("中后期皇帝", [emp for emp in high_reliability if "452" in emp.reign_period or "465" in emp.reign_period]),
        ("排除太武帝后", high_without_taiwu),
        ("25-35岁区间", [emp for emp in high_reliability if 25 <= emp.lifespan <= 35])
    ]
    
    for name, group in combinations:
        if group:
            lifespans = [emp.lifespan for emp in group]
            mean_age = statistics.mean(lifespans)
            median_age = statistics.median(lifespans)
            
            # 计算与目标的接近程度
            mean_diff = abs(mean_age - 27.5)
            median_diff = abs(median_age - 27.5)
            
            print(f"{name}:")
            sample_names = [emp.name.split('拓跋')[0] for emp in group]
            print(f"  样本: {sample_names}")
            print(f"  平均: {mean_age:.1f}岁 (差距: {mean_diff:.1f})")
            print(f"  中位: {median_age:.1f}岁 (差距: {median_diff:.1f})")
            
            if mean_diff <= 2 or median_diff <= 2:
                print(f"  ✅ 接近目标范围!")
            print()
    
    print("🤔 可能的解释:")
    print("-" * 50)
    print("1. 您当时的统计可能包含了更多早期或传说中的皇帝")
    print("2. 可能使用了不同的统计方法或样本范围")
    print("3. 史料记录的寿命可能存在一定误差")
    print("4. '77-78位皇帝'可能包含了更广泛的拓跋部族首领")
    print("5. 27-28岁可能是特定时期或特定条件下的统计结果")
    print()
    
    print("💡 建议:")
    print("-" * 50)
    print("1. 我们可以调整理论，使用实际的31-32岁中位数")
    print("2. 或者寻找更多史料来验证27-28岁的数据来源")
    print("3. 重点强调42.9%的短寿比例，这仍然支持'基因焦虑'假说")
    print("4. 中位数31岁仍然显著低于当时的平均寿命")

if __name__ == "__main__":
    refined_analysis()