knowledge_search/run_similarity_analysis.py

#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""
快速运行相似度分析脚本
"""

import os
import json
import logging
from src.analyzers.similarity_analyzer import SimilarityAnalyzer

# 配置日志
logging.basicConfig(
    level=logging.INFO,
    format='%(asctime)s - %(levelname)s - %(message)s'
)

def main():
    """主函数"""
    # 读取解构分析结果
    deconstruction_path = "output_v2/deep_analysis_results.json"

    if not os.path.exists(deconstruction_path):
        print(f"❌ 解构分析结果不存在: {deconstruction_path}")
        print("   请先运行: python3 main.py --enable-stage5 --enable-stage6")
        return

    print(f"📖 加载解构分析结果: {deconstruction_path}")
    with open(deconstruction_path, 'r', encoding='utf-8') as f:
        deconstruction_results = json.load(f)

    print(f"✓ 加载了 {len(deconstruction_results.get('results', []))} 个解构结果")

    # 创建分析器
    print("\n🚀 初始化相似度分析器...")
    analyzer = SimilarityAnalyzer(
        weight_embedding=0.5,        # 向量模型权重
        weight_semantic=0.5,         # LLM模型权重
        max_workers=5,               # 并发数
        min_similarity=0.0,          # 最小相似度阈值（0.0保留所有）
        target_features=None,        # None = 处理所有特征
        evaluation_results_path='output_v2/evaluated_results.json',
        update_evaluation_scores=True  # 自动计算综合得分P
    )

    # 运行分析
    print("\n" + "=" * 60)
    print("开始相似度分析...")
    print("=" * 60)
    similarity_results = analyzer.run(deconstruction_results)

    print("\n" + "=" * 60)
    print("✅ 相似度分析完成！")
    print("=" * 60)

    # 打印统计信息
    meta = similarity_results['metadata']
    stats = meta['overall_statistics']

    print(f"\n📊 统计结果:")
    print(f"  - 处理帖子数: {stats['total_notes']}")
    print(f"  - 提取特征总数: {stats['total_features_extracted']}")
    print(f"  - 平均特征/帖子: {stats['avg_features_per_note']}")
    print(f"  - 平均最高相似度: {stats['avg_max_similarity']}")
    print(f"  - 包含高相似度特征的帖子: {stats['notes_with_high_similarity']}")
    print(f"  - 总耗时: {meta['processing_time_seconds']}秒")

    print(f"\n📁 输出文件:")
    print(f"  - output_v2/similarity_analysis_results.json (相似度分析结果)")
    print(f"  - output_v2/evaluated_results.json (已更新综合得分P)")

    print(f"\n🎨 现在可以运行可视化:")
    print(f"  python3 src/visualizers/deconstruction_visualizer.py")

if __name__ == '__main__':
    main()