150 lines
4.7 KiB
Python
150 lines
4.7 KiB
Python
#!/usr/bin/env python3
|
|
# -*- coding: utf-8 -*-
|
|
"""
|
|
清理空对话脚本
|
|
删除 total_messages 为 0 的对话记录及其对应文件
|
|
"""
|
|
|
|
import json
|
|
import os
|
|
import shutil
|
|
from pathlib import Path
|
|
|
|
def cleanup_empty_conversations():
|
|
"""清理空对话"""
|
|
|
|
# 文件路径
|
|
base_dir = Path("/Users/jojo/Desktop/agent_kimi_源码备份/data/conversations")
|
|
index_file = base_dir / "index.json"
|
|
|
|
print(f"🔍 检查文件: {index_file}")
|
|
|
|
# 检查文件是否存在
|
|
if not index_file.exists():
|
|
print(f"❌ 错误: {index_file} 不存在")
|
|
return
|
|
|
|
# 创建备份
|
|
backup_file = base_dir / "index.json.backup"
|
|
try:
|
|
shutil.copy2(index_file, backup_file)
|
|
print(f"✅ 已创建备份: {backup_file}")
|
|
except Exception as e:
|
|
print(f"❌ 创建备份失败: {e}")
|
|
return
|
|
|
|
# 读取索引文件
|
|
try:
|
|
with open(index_file, 'r', encoding='utf-8') as f:
|
|
data = json.load(f)
|
|
except Exception as e:
|
|
print(f"❌ 读取索引文件失败: {e}")
|
|
return
|
|
|
|
print(f"📊 索引文件加载成功，共 {len(data)} 条记录")
|
|
|
|
# 查找空对话
|
|
empty_conversations = []
|
|
valid_conversations = []
|
|
|
|
for conv_id, conv_data in data.items():
|
|
if conv_data.get("total_messages", 0) == 0:
|
|
empty_conversations.append((conv_id, conv_data))
|
|
print(f"🗑️ 发现空对话: {conv_id} (创建于: {conv_data.get('created_at', '未知')})")
|
|
else:
|
|
valid_conversations.append((conv_id, conv_data))
|
|
|
|
print(f"\n📈 统计结果:")
|
|
print(f" 空对话数量: {len(empty_conversations)}")
|
|
print(f" 有效对话数量: {len(valid_conversations)}")
|
|
|
|
if len(empty_conversations) == 0:
|
|
print("✨ 没有发现空对话，无需清理")
|
|
return
|
|
|
|
# 确认删除
|
|
print(f"\n⚠️ 将要删除 {len(empty_conversations)} 个空对话:")
|
|
for conv_id, conv_data in empty_conversations[:5]: # 只显示前5个
|
|
title = conv_data.get('title', '无标题')
|
|
created = conv_data.get('created_at', '未知时间')
|
|
print(f" - {conv_id}: {title} ({created})")
|
|
|
|
if len(empty_conversations) > 5:
|
|
print(f" ... 还有 {len(empty_conversations) - 5} 个")
|
|
|
|
confirm = input("\n❓ 确认删除这些空对话吗? (y/N): ").strip().lower()
|
|
if confirm not in ['y', 'yes']:
|
|
print("❌ 操作已取消")
|
|
return
|
|
|
|
# 删除对话文件
|
|
deleted_files = []
|
|
failed_deletions = []
|
|
|
|
for conv_id, conv_data in empty_conversations:
|
|
conv_file = base_dir / f"{conv_id}.json"
|
|
|
|
try:
|
|
if conv_file.exists():
|
|
conv_file.unlink() # 删除文件
|
|
deleted_files.append(conv_file.name)
|
|
print(f"✅ 删除文件: {conv_file.name}")
|
|
else:
|
|
print(f"⚠️ 文件不存在: {conv_file.name}")
|
|
except Exception as e:
|
|
failed_deletions.append((conv_file.name, str(e)))
|
|
print(f"❌ 删除失败: {conv_file.name} - {e}")
|
|
|
|
# 更新索引文件
|
|
try:
|
|
new_data = {conv_id: conv_data for conv_id, conv_data in valid_conversations}
|
|
|
|
with open(index_file, 'w', encoding='utf-8') as f:
|
|
json.dump(new_data, f, ensure_ascii=False, indent=2)
|
|
|
|
print(f"✅ 索引文件已更新，保留 {len(new_data)} 条记录")
|
|
|
|
except Exception as e:
|
|
print(f"❌ 更新索引文件失败: {e}")
|
|
print("🔄 正在恢复备份...")
|
|
try:
|
|
shutil.copy2(backup_file, index_file)
|
|
print("✅ 备份已恢复")
|
|
except Exception as e2:
|
|
print(f"❌ 恢复备份失败: {e2}")
|
|
return
|
|
|
|
# 总结
|
|
print(f"\n🎉 清理完成!")
|
|
print(f" 删除的对话记录: {len(empty_conversations)} 个")
|
|
print(f" 删除的文件: {len(deleted_files)} 个")
|
|
|
|
if failed_deletions:
|
|
print(f" 删除失败: {len(failed_deletions)} 个")
|
|
for filename, error in failed_deletions:
|
|
print(f" - {filename}: {error}")
|
|
|
|
print(f" 备份文件: {backup_file}")
|
|
print(f" 剩余有效对话: {len(valid_conversations)} 个")
|
|
|
|
def main():
|
|
"""主函数"""
|
|
print("=" * 50)
|
|
print("🧹 空对话清理工具")
|
|
print("=" * 50)
|
|
|
|
try:
|
|
cleanup_empty_conversations()
|
|
except KeyboardInterrupt:
|
|
print("\n❌ 操作被用户取消")
|
|
except Exception as e:
|
|
print(f"\n💥 发生未预期的错误: {e}")
|
|
import traceback
|
|
traceback.print_exc()
|
|
|
|
print("\n" + "=" * 50)
|
|
print("🏁 脚本结束")
|
|
|
|
if __name__ == "__main__":
|
|
main()
|