ai_Image_review/stats_similarity.py

# -*- coding: utf-8 -*-
"""
图片相似度统计脚本
统计各状态的图片数量和重复率
"""

import configparser
import pymysql
from datetime import datetime


def main():
    # 读取配置
    config = configparser.ConfigParser()
    config.read('config.ini', encoding='utf-8')
    
    # 连接数据库
    db_conn = pymysql.connect(
        host=config.get('database', 'host'),
        port=config.getint('database', 'port'),
        user=config.get('database', 'user'),
        password=config.get('database', 'password'),
        database=config.get('database', 'database'),
        charset=config.get('database', 'charset'),
        cursorclass=pymysql.cursors.DictCursor
    )
    
    print("=" * 70)
    print(f"图片相似度统计报告 - {datetime.now().strftime('%Y-%m-%d %H:%M:%S')}")
    print("=" * 70)
    
    with db_conn.cursor() as cursor:
        # 1. 总体统计
        print("\n【一、总体统计】\n")
        
        stats = [
            ("待处理 (draft/draft)", "status = 'draft' AND similarity = 'draft'"),
            ("重复图片 (similarity/yes)", "status = 'similarity' AND similarity = 'yes'"),
            ("不重复图片 (tag_extension/calc)", "status = 'tag_extension' AND similarity = 'calc'"),
            ("待重算 (draft/recalc)", "status = 'draft' AND similarity = 'recalc'"),
        ]
        
        results = {}
        for name, condition in stats:
            cursor.execute(f"SELECT COUNT(*) as cnt FROM ai_image_tags WHERE {condition}")
            count = cursor.fetchone()['cnt']
            results[name] = count
            print(f"  {name:<35} : {count:>10,} 张")
        
        # 总计
        cursor.execute("SELECT COUNT(*) as cnt FROM ai_image_tags")
        total = cursor.fetchone()['cnt']
        print(f"  {'─' * 50}")
        print(f"  {'总计':<35} : {total:>10,} 张")
        
        # 2. 重复率计算
        print("\n【二、重复率分析】\n")
        
        duplicate_count = results["重复图片 (similarity/yes)"]
        unique_count = results["不重复图片 (tag_extension/calc)"]
        processed = duplicate_count + unique_count
        
        if processed > 0:
            duplicate_rate = (duplicate_count / processed) * 100
            print(f"  已处理图片数    : {processed:>10,} 张")
            print(f"  重复图片数      : {duplicate_count:>10,} 张")
            print(f"  不重复图片数    : {unique_count:>10,} 张")
            print(f"  重复率          : {duplicate_rate:>10.2f} %")
        else:
            print("  暂无已处理的图片数据")
        
        # 3. 处理进度
        print("\n【三、处理进度】\n")
        
        pending = results["待处理 (draft/draft)"]
        recalc = results["待重算 (draft/recalc)"]
        
        if total > 0:
            progress = (processed / total) * 100
            print(f"  总进度          : {progress:>10.2f} %")
            print(f"  待处理          : {pending:>10,} 张")
            print(f"  待重算          : {recalc:>10,} 张")
            print(f"  已完成          : {processed:>10,} 张")
        
        # 4. 相似度分数分布（仅重复图片）
        print("\n【四、相似度分数分布】\n")
        
        cursor.execute("""
            SELECT 
                CASE 
                    WHEN similarity_score >= 0.99 THEN '0.99-1.00 (几乎相同)'
                    WHEN similarity_score >= 0.97 THEN '0.97-0.99 (非常相似)'
                    WHEN similarity_score >= 0.95 THEN '0.95-0.97 (高度相似)'
                    WHEN similarity_score >= 0.94 THEN '0.94-0.95 (相似)'
                    ELSE '< 0.94 (其他)'
                END as score_range,
                COUNT(*) as cnt
            FROM ai_image_tags
            WHERE status = 'similarity' AND similarity = 'yes'
            GROUP BY score_range
            ORDER BY score_range DESC
        """)
        
        score_stats = cursor.fetchall()
        if score_stats:
            for row in score_stats:
                print(f"  {row['score_range']:<25} : {row['cnt']:>10,} 张")
        else:
            print("  暂无重复图片数据")
        
        # 5. 最近处理记录
        print("\n【五、最近 10 条重复记录】\n")
        
        cursor.execute("""
            SELECT id, image_name, similarity_image_tags_id, similarity_score, updated_at
            FROM ai_image_tags
            WHERE status = 'similarity' AND similarity = 'yes'
            ORDER BY updated_at DESC
            LIMIT 10
        """)
        
        recent = cursor.fetchall()
        if recent:
            print(f"  {'ID':<10} {'相似ID':<10} {'分数':<10} {'更新时间':<20}")
            print(f"  {'-' * 55}")
            for row in recent:
                score = f"{row['similarity_score']:.4f}" if row['similarity_score'] else "-"
                updated = row['updated_at'].strftime('%Y-%m-%d %H:%M') if row['updated_at'] else "-"
                print(f"  {row['id']:<10} {row['similarity_image_tags_id']:<10} {score:<10} {updated:<20}")
        else:
            print("  暂无重复记录")
    
    db_conn.close()
    
    print("\n" + "=" * 70)
    print("统计完成")
    print("=" * 70)


if __name__ == '__main__':
    main()