chore: add ktyyds bot script

bots/ktyyds_bot/finance_agent.py

@@ -0,0 +1,909 @@
+#!/usr/bin/env python3
+"""
+财务管家 AI Agent - 终极版
+真正智能、极致体验
+创作者：kt暴君
+版本：3.0 Ultimate
+"""
+
+import asyncio
+import json
+import logging
+import shutil
+from typing import Dict, Any, List, Optional, Tuple, Deque
+from datetime import datetime, timedelta
+import subprocess
+import os
+import re
+from collections import defaultdict, Counter, deque
+
+logger = logging.getLogger(__name__)
+
+class FinanceAgentUltimate:
+    """财务管家 - 终极智能版"""
+    
+    def __init__(self):
+        self.name = "财务管家"
+        self.creator = "kt暴君"
+        self.version = "3.0 Ultimate"
+        self.api_url = "http://192.168.9.149:5666/api/finance/transactions"
+        
+        # 缓存数据，避免重复请求
+        self.data_cache = None
+        self.cache_time = None
+        self.cache_ttl = 60  # 缓存60秒
+
+        # 上下文记忆系统（按用户ID存储）
+        self.user_contexts = {}  # {user_id: {state, data, timestamp}}
+        self.general_chat_cli = os.getenv("FINANCE_AGENT_GENERAL_CLI", "codex")
+        self.general_histories: Dict[str, Deque[Tuple[str, str]]] = {}
+        
+    async def initialize(self):
+        """初始化"""
+        logger.info(f"✅ {self.name} v{self.version} 已启动 - 终极智能模式")
+        # 预加载数据
+        await self._get_data(force_refresh=True)
+        return True
+    
+    async def _get_data(self, force_refresh=False) -> List[Dict]:
+        """获取数据（带缓存）"""
+        now = datetime.now()
+
+        if (
+            not force_refresh
+            and self.data_cache is not None
+            and self.cache_time is not None
+            and (now - self.cache_time).total_seconds() < self.cache_ttl
+        ):
+            return self.data_cache
+
+        try:
+            import requests
+
+            resp = requests.get(self.api_url, timeout=10)
+            if resp.status_code == 200:
+                payload = resp.json()
+                if isinstance(payload, dict):
+                    self.data_cache = payload.get("data", [])
+                else:
+                    self.data_cache = []
+                self.cache_time = now
+                logger.info(f"数据已更新: {len(self.data_cache)} 条记录")
+                return self.data_cache
+
+            logger.error(f"获取数据失败: 状态码 {resp.status_code}")
+        except Exception as e:
+            logger.error(f"获取数据失败: {e}")
+
+        return self.data_cache or []
+    
+    async def process_command(self, command: str, context: Dict = None) -> str:
+        """主处理入口 - 智能路由（支持上下文）"""
+        cmd = command.lower().strip()
+        user_id = context.get("user_id", "default") if context else "default"
+        
+        # 检查是否有待处理的上下文
+        if user_id in self.user_contexts:
+            ctx = self.user_contexts[user_id]
+            # 检查上下文是否过期（5分钟）
+            from datetime import datetime, timedelta
+            if datetime.now() - ctx.get("timestamp", datetime.now()) < timedelta(minutes=5):
+                # 尝试处理上下文响应
+                result = await self._handle_context_response(command, user_id, ctx)
+                if result:
+                    return result
+            else:
+                # 上下文过期，清除
+                del self.user_contexts[user_id]
+        
+        # 1. 账单查询（最高优先级）
+        if self._is_bill_query(command):
+            return await self._handle_bill_query(command, context)
+        
+        # 2. 网站相关（优先于状态检查，避免被拦截）
+        if any(kw in cmd for kw in ["网站", "website", "web"]) and "状态" in cmd:
+            return await self._handle_website(command)
+        
+        # 3. 虚拟机/服务器（优先处理具体对象）
+        if any(kw in cmd for kw in ["虚拟机", "vm", "服务器"]) and not any(kw in cmd for kw in ["状态", "重启"]):
+            return await self._handle_vm(command)
+        
+        # 4. 状态检查
+        if any(kw in cmd for kw in ["状态", "status", "运行", "健康", "正常吗"]):
+            return await self._handle_status_check(command)
+        
+        # 5. 统计分析
+        if any(kw in cmd for kw in ["统计", "分析", "趋势", "汇总", "总共"]):
+            return await self._handle_statistics(command)
+        
+        # 6. 数据查询
+        if any(kw in cmd for kw in ["本月", "今天", "本周", "昨天", "最近"]):
+            return await self._handle_time_query(command)
+        
+        # 7. 系统管理
+        if "重启" in cmd:
+            return await self._handle_restart(command)
+        
+        # 8. 日志查看
+        if any(kw in cmd for kw in ["日志", "log", "错误", "异常"]):
+            return await self._handle_logs(command)
+        
+        # 9. 帮助
+        if any(kw in cmd for kw in ["帮助", "help", "功能", "怎么用", "命令"]):
+            return await self._handle_help(command)
+        
+        # 10. 智能对话
+        return await self._handle_conversation(command, context)
+    
+    # ========== 上下文管理 ==========
+    
+    async def _handle_context_response(self, command: str, user_id: str, ctx: Dict) -> Optional[str]:
+        """处理上下文响应"""
+        state = ctx.get("state")
+        
+        # 年份选择上下文
+        if state == "awaiting_year_selection":
+            # 检查用户是否输入了年份
+            year_match = re.search(r'20(2[0-9])', command)
+            if year_match:
+                year = int(year_match.group(0))
+                months = ctx.get("months", [])
+                
+                # 更新月份的年份
+                for m in months:
+                    m["year"] = year
+                
+                # 清除上下文
+                del self.user_contexts[user_id]
+                
+                # 执行查询
+                data = await self._get_data()
+                return await self._execute_bill_query(data, months)
+            
+            # 检查是否输入了年份关键词
+            if "2024" in command or "2025" in command or "今年" in command or "去年" in command:
+                if "2024" in command:
+                    year = 2024
+                elif "2025" in command or "今年" in command:
+                    year = 2025
+                elif "去年" in command:
+                    year = datetime.now().year - 1
+                else:
+                    return None
+                
+                months = ctx.get("months", [])
+                for m in months:
+                    m["year"] = year
+                
+                del self.user_contexts[user_id]
+                data = await self._get_data()
+                return await self._execute_bill_query(data, months)
+        
+        return None
+    
+    def _set_context(self, user_id: str, state: str, data: Dict):
+        """设置用户上下文"""
+        from datetime import datetime
+        self.user_contexts[user_id] = {
+            "state": state,
+            "timestamp": datetime.now(),
+            **data
+        }
+    
+    # ========== 账单查询 ==========
+    
+    def _is_bill_query(self, cmd: str) -> bool:
+        """判断是否是账单查询"""
+        month_patterns = [r'\d{1,2}月', r'[一二三四五六七八九十]{1,2}月', r'\d{4}年\d{1,2}月']
+        has_month = any(re.search(p, cmd) for p in month_patterns)
+        has_query = any(w in cmd for w in ["账单", "记录", "查", "看", "显示"])
+        return has_month and has_query
+    
+    async def _handle_bill_query(self, command: str, context: Dict = None) -> str:
+        """处理账单查询 - 极致智能（支持上下文）"""
+        data = await self._get_data()
+        user_id = context.get("user_id", "default") if context else "default"
+        
+        # 解析月份
+        months = self._parse_months(command)
+        if not months:
+            return "❓ 请告诉我要查询哪个月份\n\n例如：\n• 查9月账单\n• 查2024年9月\n• 查7月和8月"
+        
+        # 检查是否需要询问年份（只有当用户使用当前年份且命令中没有明确年份时才询问）
+        current_year = datetime.now().year
+        # 检查命令中是否包含年份数字
+        has_explicit_year = bool(re.search(r'\d{4}年', command))
+        need_confirm = any(m["year"] == current_year for m in months) and not has_explicit_year
+        
+        if need_confirm:
+            # 检查历史数据
+            year_stats = self._get_year_stats(data, months)
+            if len(year_stats) > 1:
+                # 多年都有数据，需要用户确认
+                # 保存上下文
+                self._set_context(user_id, "awaiting_year_selection", {
+                    "months": months,
+                    "year_stats": year_stats,
+                    "original_command": command
+                })
+                
+                month_names = "、".join([f"{m['month']}月" for m in months])
+                options = []
+                for year in sorted(year_stats.keys(), reverse=True):
+                    options.append(f"• {year}年{month_names}: {year_stats[year]['count']}条记录，总计 ¥{year_stats[year]['total']:.3f}")
+                
+                return f"📊 发现多个年份都有{month_names}的数据：\n\n" + "\n".join(options) + f"\n\n❓ 请直接回复年份：\n• 2025\n• 2024"
+        
+        # 执行查询
+        return await self._execute_bill_query(data, months)
+    
+    def _parse_months(self, cmd: str) -> List[Dict]:
+        """解析月份"""
+        months = []
+        current_year = datetime.now().year
+        
+        cn_map = {"一": 1, "二": 2, "三": 3, "四": 4, "五": 5, "六": 6,
+                  "七": 7, "八": 8, "九": 9, "十": 10, "十一": 11, "十二": 12}
+        
+        # 完整格式：2024年9月
+        for m in re.finditer(r'(\d{4})年(\d{1,2})月', cmd):
+            year, month = int(m.group(1)), int(m.group(2))
+            if 1 <= month <= 12 and 2020 <= year <= 2030:
+                months.append({"year": year, "month": month})
+        
+        # 简单格式：9月（使用当前年份）
+        if not months:
+            for m in re.finditer(r'(\d{1,2})月', cmd):
+                month = int(m.group(1))
+                if 1 <= month <= 12:
+                    months.append({"year": current_year, "month": month})
+            
+            # 中文月份
+            for cn, num in cn_map.items():
+                if f"{cn}月" in cmd and not any(m["month"] == num for m in months):
+                    months.append({"year": current_year, "month": num})
+        
+        # 去重
+        seen = set()
+        unique = []
+        for m in months:
+            key = (m["year"], m["month"])
+            if key not in seen:
+                seen.add(key)
+                unique.append(m)
+        
+        return unique
+    
+    def _get_year_stats(self, data: List[Dict], months: List[Dict]) -> Dict:
+        """获取各年份的统计"""
+        stats = {}
+        for item in data:
+            date_str = item.get("transactionDate", "")
+            if not date_str:
+                continue
+            
+            try:
+                date = datetime.strptime(date_str, "%Y-%m-%d")
+                for m in months:
+                    if date.month == m["month"]:
+                        year = date.year
+                        if year not in stats:
+                            stats[year] = {"count": 0, "total": 0}
+                        stats[year]["count"] += 1
+                        stats[year]["total"] += float(item.get("amount", 0))
+                        break
+            except:
+                continue
+        
+        return stats
+    
+    async def _execute_bill_query(self, data: List[Dict], months: List[Dict]) -> str:
+        """执行账单查询"""
+        # 筛选数据
+        filtered = []
+        for item in data:
+            date_str = item.get("transactionDate", "")
+            if not date_str:
+                continue
+            
+            try:
+                date = datetime.strptime(date_str, "%Y-%m-%d")
+                for m in months:
+                    if date.year == m["year"] and date.month == m["month"]:
+                        filtered.append(item)
+                        break
+            except:
+                continue
+        
+        if not filtered:
+            month_desc = " + ".join([f"{m['year']}年{m['month']}月" for m in months])
+            # 智能提示有数据的月份
+            all_months = set()
+            for item in data:
+                date_str = item.get("transactionDate", "")
+                if date_str:
+                    try:
+                        all_months.add(date_str[:7])
+                    except:
+                        pass
+            recent = sorted(all_months, reverse=True)[:5]
+            hints = "\n".join([f"• {m}" for m in recent])
+            return f"📊 {month_desc} 暂无记录\n\n💡 最近有数据的月份：\n{hints}"
+        
+        # 生成详细报告
+        filtered.sort(key=lambda x: x.get("transactionDate", ""), reverse=True)
+        
+        # 分月统计
+        by_month = defaultdict(list)
+        for item in filtered:
+            month_key = item.get("transactionDate", "")[:7]
+            by_month[month_key].append(item)
+        
+        # 生成报告
+        total = sum(float(i.get("amount", 0)) for i in filtered)
+        month_desc = " + ".join([f"{m['year']}年{m['month']}月" for m in months])
+        
+        report = f"📊 {month_desc} 账单详情\n{'='*50}\n\n"
+        
+        for month_key in sorted(by_month.keys(), reverse=True):
+            items = by_month[month_key]
+            subtotal = sum(float(i.get("amount", 0)) for i in items)
+            
+            # 格式化月份
+            y, m = month_key.split('-')
+            report += f"📅 {y}年{int(m)}月 ({len(items)}笔)\n{'-'*60}\n"
+            
+            # 表格头
+            report += f"{'序号':<4} {'日期':<12} {'金额':<12} {'币种':<6} {'描述'}\n"
+            report += f"{'━'*60}\n"
+            
+            # 显示所有记录（完整）
+            for idx, item in enumerate(items, 1):
+                date = item.get("transactionDate", "")
+                amt = float(item.get("amount", 0))
+                desc = item.get("description", "无描述")
+                curr = item.get("currency", "CNY")
+                
+                # 截断过长描述
+                if len(desc) > 25:
+                    desc = desc[:22] + "..."
+                
+                report += f"{idx:<4} {date:<12} {amt:<12.2f} {curr:<6} {desc}\n"
+            
+            report += f"{'-'*60}\n"
+            report += f"💰 小计: ¥{subtotal:.3f}\n\n"
+        
+        report += f"{'='*50}\n"
+        report += f"💰 总计: ¥{total:.3f}\n"
+        report += f"📝 共 {len(filtered)} 笔交易\n"
+        report += f"🕐 {datetime.now().strftime('%Y-%m-%d %H:%M:%S')}\n"
+        
+        return report
+    
+    # ========== 状态检查 ==========
+    
+    async def _handle_status_check(self, command: str) -> str:
+        """状态检查 - 详细健康报告"""
+        checks = []
+        
+        # 1. 机器人状态
+        bot_status = await self._check_bot_status()
+        checks.append(bot_status)
+        
+        # 2. API状态
+        api_status = await self._check_api_status()
+        checks.append(api_status)
+        
+        # 3. 虚拟机状态
+        vm_status = await self._check_vm_status()
+        checks.append(vm_status)
+        
+        # 4. 数据统计
+        data_status = await self._check_data_status()
+        checks.append(data_status)
+        
+        # 汇总健康度
+        health_score = sum(1 for c in checks if "✅" in c["status"]) / len(checks) * 100
+        
+        report = f"📊 系统健康度报告\n{'='*50}\n\n"
+        report += f"🏥 健康评分: {health_score:.0f}%\n\n"
+        
+        for check in checks:
+            report += f"{check['icon']} {check['name']}: {check['status']}\n"
+            if check.get("detail"):
+                report += f"   {check['detail']}\n"
+        
+        report += f"\n{'='*50}\n"
+        report += f"🕐 {datetime.now().strftime('%Y-%m-%d %H:%M:%S')}\n"
+        
+        if health_score < 80:
+            report += "\n⚠️ 系统存在问题，建议检查！"
+        
+        return report
+    
+    async def _check_bot_status(self) -> Dict:
+        """检查机器人状态"""
+        try:
+            result = subprocess.run(
+                "ps aux | grep telegram_webhook_bot.py | grep -v grep",
+                shell=True, capture_output=True, text=True, timeout=5
+            )
+            running = bool(result.stdout.strip())
+            
+            if running:
+                # 获取运行时间
+                lines = result.stdout.strip().split('\n')
+                if lines:
+                    return {
+                        "icon": "🤖", "name": "机器人",
+                        "status": "✅ 运行中",
+                        "detail": "服务正常"
+                    }
+            
+            return {"icon": "🤖", "name": "机器人", "status": "❌ 已停止", "detail": "需要重启"}
+        except:
+            return {"icon": "🤖", "name": "机器人", "status": "⚠️ 检查失败", "detail": None}
+    
+    async def _check_api_status(self) -> Dict:
+        """检查API状态"""
+        try:
+            data = await self._get_data()
+            count = len(data)
+            return {
+                "icon": "🌐", "name": "API服务",
+                "status": f"✅ 正常",
+                "detail": f"{count}条记录"
+            }
+        except:
+            return {"icon": "🌐", "name": "API服务", "status": "❌ 异常", "detail": "无法连接"}
+    
+    async def _check_vm_status(self) -> Dict:
+        """检查虚拟机状态"""
+        try:
+            import socket
+            sock = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
+            sock.settimeout(3)
+            result = sock.connect_ex(("192.168.9.149", 22))
+            sock.close()
+            
+            if result == 0:
+                return {"icon": "💻", "name": "虚拟机", "status": "✅ 在线", "detail": "SSH可达"}
+            return {"icon": "💻", "name": "虚拟机", "status": "❌ 离线", "detail": "无法连接"}
+        except:
+            return {"icon": "💻", "name": "虚拟机", "status": "⚠️ 检查失败", "detail": None}
+    
+    async def _check_data_status(self) -> Dict:
+        """检查数据状态"""
+        try:
+            data = await self._get_data()
+            if not data:
+                return {"icon": "📊", "name": "数据", "status": "⚠️ 无数据", "detail": None}
+            
+            # 最新记录时间
+            dates = [d.get("transactionDate", "") for d in data if d.get("transactionDate")]
+            if dates:
+                latest = max(dates)
+                return {
+                    "icon": "📊", "name": "数据",
+                    "status": "✅ 正常",
+                    "detail": f"最新: {latest}"
+                }
+            
+            return {"icon": "📊", "name": "数据", "status": "⚠️ 异常", "detail": "无日期"}
+        except:
+            return {"icon": "📊", "name": "数据", "status": "❌ 错误", "detail": None}
+    
+    # ========== 统计分析 ==========
+    
+    async def _handle_statistics(self, command: str) -> str:
+        """统计分析 - 可视化报告"""
+        data = await self._get_data()
+        
+        if not data:
+            return "❌ 暂无数据"
+        
+        now = datetime.now()
+        today = now.date()
+        month_start = now.replace(day=1).date()
+        year_start = now.replace(month=1, day=1).date()
+        
+        today_total = 0
+        month_total = 0
+        year_total = 0
+        
+        # 分类统计
+        categories = Counter()
+        monthly = defaultdict(float)
+        
+        for item in data:
+            date_str = item.get("transactionDate", "")
+            if not date_str:
+                continue
+            
+            try:
+                date = datetime.strptime(date_str, "%Y-%m-%d").date()
+                amount = float(item.get("amount", 0))
+                
+                if date == today:
+                    today_total += amount
+                if date >= month_start:
+                    month_total += amount
+                if date >= year_start:
+                    year_total += amount
+                
+                # 月度统计
+                month_key = date_str[:7]
+                monthly[month_key] += amount
+                
+                # 分类统计
+                desc = item.get("description", "其他")
+                categories[desc] += amount
+            except:
+                continue
+        
+        # 生成报告
+        report = f"📊 财务统计分析\n{'='*50}\n\n"
+        
+        # 基础统计
+        report += "📈 基础数据\n"
+        report += f"• 今日支出: ¥{today_total:.3f}\n"
+        report += f"• 本月支出: ¥{month_total:.3f}\n"
+        report += f"• 本年支出: ¥{year_total:.3f}\n"
+        report += f"• 总记录数: {len(data)} 条\n\n"
+        
+        # 月度趋势（最近6个月）
+        recent_months = sorted(monthly.keys(), reverse=True)[:6]
+        if recent_months:
+            report += "📅 月度趋势（最近6个月）\n"
+            max_amt = max(monthly[m] for m in recent_months)
+            for month in recent_months:
+                amt = monthly[month]
+                bar_len = int((amt / max_amt) * 20) if max_amt > 0 else 0
+                bar = "█" * bar_len
+                report += f"{month}: ¥{amt:>10.2f} {bar}\n"
+            report += "\n"
+        
+        # Top消费
+        top_cat = categories.most_common(5)
+        if top_cat:
+            report += "🏆 Top 5 消费项目\n"
+            for idx, (cat, amt) in enumerate(top_cat, 1):
+                report += f"{idx}. {cat[:20]:20} ¥{amt:.3f}\n"
+        
+        report += f"\n{'='*50}\n"
+        report += f"🕐 {datetime.now().strftime('%Y-%m-%d %H:%M:%S')}\n"
+        
+        return report
+    
+    # ========== 其他功能 ==========
+    
+    async def _handle_time_query(self, command: str) -> str:
+        """时间范围查询"""
+        data = await self._get_data()
+        now = datetime.now()
+        
+        # 解析时间范围
+        if "今天" in command or "今日" in command:
+            start = now.date()
+            title = "今日"
+        elif "昨天" in command:
+            start = (now - timedelta(days=1)).date()
+            title = "昨日"
+        elif "本周" in command:
+            start = (now - timedelta(days=now.weekday())).date()
+            title = "本周"
+        elif "本月" in command:
+            start = now.replace(day=1).date()
+            title = "本月"
+        else:
+            start = (now - timedelta(days=7)).date()
+            title = "最近7天"
+        
+        # 筛选数据
+        filtered = []
+        for item in data:
+            date_str = item.get("transactionDate", "")
+            if date_str:
+                try:
+                    date = datetime.strptime(date_str, "%Y-%m-%d").date()
+                    if date >= start:
+                        filtered.append(item)
+                except:
+                    pass
+        
+        if not filtered:
+            return f"📊 {title}暂无交易记录"
+        
+        total = sum(float(i.get("amount", 0)) for i in filtered)
+        
+        report = f"📊 {title}交易汇总\n{'='*50}\n\n"
+        report += f"💰 总金额: ¥{total:.3f}\n"
+        report += f"📝 交易笔数: {len(filtered)}\n\n"
+        
+        # 最近5笔
+        filtered.sort(key=lambda x: x.get("transactionDate", ""), reverse=True)
+        report += "最近交易:\n"
+        for idx, item in enumerate(filtered[:5], 1):
+            date = item.get("transactionDate", "")
+            amt = item.get("amount", 0)
+            desc = item.get("description", "")[:20]
+            report += f"{idx}. {date} ¥{amt} {desc}\n"
+        
+        if len(filtered) > 5:
+            report += f"... 还有 {len(filtered)-5} 笔\n"
+        
+        return report
+    
+    async def _handle_restart(self, command: str) -> str:
+        """重启服务"""
+        cmd_lower = command.lower()
+        
+        # 如果是询问怎么做，提供指导而不是直接执行
+        if any(w in cmd_lower for w in ["怎么", "如何", "怎样", "如何做"]):
+            return """
+🔧 重启服务指南
+
+【重启机器人】
+直接输入：重启机器人
+
+【重启网站】
+直接输入：重启网站
+
+⚠️ 注意：重启会导致短暂服务中断
+"""
+        
+        # 明确的重启命令才执行
+        if "机器人" in command or "bot" in command:
+            return await self._restart_bot()
+        
+        return "请明确要重启什么：\n• 重启机器人\n• 重启网站"
+    
+    async def _restart_bot(self) -> str:
+        """重启机器人"""
+        try:
+            subprocess.run("pkill -f telegram_webhook_bot.py", shell=True)
+            await asyncio.sleep(1)
+            subprocess.Popen(
+                ["python3", "/Users/fuwuqi/telegram_webhook_bot.py"],
+                stdout=open("/Users/fuwuqi/bot.log", "a"),
+                stderr=subprocess.STDOUT,
+                start_new_session=True
+            )
+            await asyncio.sleep(2)
+            result = subprocess.run(
+                "ps aux | grep telegram_webhook_bot.py | grep -v grep | wc -l",
+                shell=True, capture_output=True, text=True
+            )
+            if int(result.stdout.strip()) > 0:
+                return "✅ 机器人重启成功！"
+            return "❌ 重启失败，请检查日志"
+        except Exception as e:
+            return f"❌ 重启失败: {e}"
+    
+    async def _handle_logs(self, command: str) -> str:
+        """日志查看"""
+        try:
+            lines = 20
+            if "错误" in command or "error" in command.lower():
+                # 只显示错误
+                result = subprocess.run(
+                    f"tail -100 /Users/fuwuqi/bot.log | grep -i error | tail -20",
+                    shell=True, capture_output=True, text=True
+                )
+                content = result.stdout or "无错误日志"
+                return f"📝 错误日志\n{'='*50}\n{content}"
+            
+            with open("/Users/fuwuqi/bot.log", "r") as f:
+                content = "".join(f.readlines()[-lines:])
+            
+            return f"📝 最近日志（{lines}行）\n{'='*50}\n{content}"
+        except:
+            return "❌ 无法读取日志"
+    
+    async def _handle_help(self, command: str) -> str:
+        """帮助"""
+        return f"""
+🤖 {self.name} v{self.version}
+创作者：{self.creator}
+
+💡 我能做什么：
+
+【📊 账单查询】
+• 查9月账单 - 智能询问年份
+• 查2024年9月 - 指定年份
+• 查7月和8月 - 多月汇总
+
+【📈 统计分析】
+• 统计 - 详细财务报告
+• 本月 / 今天 - 时间范围查询
+• 分析 - 趋势和可视化
+
+【🔧 系统管理】
+• 状态 - 健康度报告
+• 重启机器人 - 安全重启
+• 日志 - 查看运行日志
+
+【💬 智能对话】
+直接说话即可，我会理解！
+
+📱 系统信息：
+• 网站: http://192.168.9.149:5666
+• 机器人: @ktyyds_bot
+• 作者: {self.creator}
+"""
+    
+    async def _handle_vm(self, command: str) -> str:
+        """虚拟机信息"""
+        return """
+💻 虚拟机信息
+{'='*50}
+• IP: 192.168.9.149
+• 用户: atai
+• 端口: 5666
+• SSH: 22
+• 数据库: /home/atai/finwise-pro/apps/backend/storage/finance.db
+"""
+    
+    async def _handle_website(self, command: str) -> str:
+        """网站相关"""
+        try:
+            import socket
+            sock = socket.socket()
+            sock.settimeout(3)
+            result = sock.connect_ex(("192.168.9.149", 5666))
+            sock.close()
+            
+            if result == 0:
+                return "🌐 网站状态: ✅ 正常\n地址: http://192.168.9.149:5666"
+            return "🌐 网站状态: ❌ 无法访问\n\n建议:\n• 检查网络\n• 检查虚拟机"
+        except:
+            return "🌐 网站状态: ⚠️ 检查失败"
+    
+    async def _handle_conversation(self, command: str, context: Dict = None) -> str:
+        """智能对话"""
+        user_id = context.get("user_id", "default") if context else "default"
+        cmd = command.lower()
+
+        if any(w in cmd for w in ["你好", "hi", "hello", "嗨"]):
+            response = f"您好！我是{self.name}，智能财务助手。\n\n输入'帮助'查看我能做什么！"
+            self._remember_exchange(user_id, command, response)
+            return response
+
+        if any(w in cmd for w in ["谢谢", "thank"]):
+            response = "不客气！随时为您服务 😊"
+            self._remember_exchange(user_id, command, response)
+            return response
+
+        general_reply = await self._generate_general_reply(user_id, command)
+        if general_reply:
+            return general_reply
+
+        if any(symbol in command for symbol in ["?", "？", "吗"]):
+            response = (
+                "我暂时没抓住您的重点，能再具体描述一下吗？\n"
+                "我擅长：账单查询、网站/虚拟机/机器人状态、财务统计，还有日常聊天。"
+            )
+            self._remember_exchange(user_id, command, response)
+            return response
+
+        response = (
+            f"🤔 我理解您说：\"{command}\"\n\n"
+            "如果是系统相关需求，可以告诉我：账单查询、网站/虚拟机状态、机器人、统计分析等。\n"
+            "如果只是想聊聊别的，也可以直接说，我会用自然语言和您交流。"
+        )
+        self._remember_exchange(user_id, command, response)
+        return response
+
+    async def _generate_general_reply(self, user_id: str, message: str) -> Optional[str]:
+        """调用外部对话模型，处理泛化对话"""
+        if not self.general_chat_cli:
+            return None
+
+        prompt = self._build_general_prompt(user_id, message)
+        reply = await self._call_general_cli(prompt)
+
+        if reply:
+            self._remember_exchange(user_id, message, reply)
+            return reply
+        return None
+
+    def _build_general_prompt(self, user_id: str, message: str) -> str:
+        """构建与Codex/Claude CLI的对话提示"""
+        history = self._get_history(user_id)
+        recent = list(history)[-6:]
+
+        formatted_history = []
+        for role, text in recent:
+            speaker = "用户" if role == "user" else "助手"
+            formatted_history.append(f"{speaker}: {text}")
+
+        history_block = "\n".join(formatted_history) if formatted_history else "（无历史对话）"
+
+        return (
+            "你是“财务管家助手”，负责财务系统网站、虚拟机和机器人运维，"
+            "也可以跟用户就任何话题自然聊天，提供常识、建议或情绪支持。\n"
+            "- 语气自然亲切，像同事聊天\n"
+            "- 回答保持务实，最多6句中文\n"
+            "- 如果问题超出掌握范围，要坦诚说明并给出下一步建议\n"
+            "- 优先考虑是否与财务系统、虚拟机、机器人相关，如是则结合实际经验回复\n"
+            "- 如果用户只是闲聊，就顺着话题交流\n\n"
+            f"历史对话：\n{history_block}\n\n"
+            f"用户: {message}\n"
+            "助手："
+        )
+
+    async def _call_general_cli(self, prompt: str) -> Optional[str]:
+        """调用本地CLI进行对话"""
+        if not shutil.which(self.general_chat_cli):
+            logger.warning(f"未找到对话CLI：{self.general_chat_cli}")
+            return None
+
+        def _run() -> Optional[str]:
+            try:
+                completed = subprocess.run(
+                    [self.general_chat_cli, prompt],
+                    capture_output=True,
+                    text=True,
+                    timeout=45
+                )
+                if completed.returncode != 0:
+                    logger.error(
+                        "对话CLI执行失败: returncode=%s, stderr=%s",
+                        completed.returncode,
+                        completed.stderr.strip()
+                    )
+                    return None
+                return self._sanitize_cli_output(completed.stdout)
+            except subprocess.TimeoutExpired:
+                logger.error("对话CLI执行超时")
+                return None
+            except Exception as exc:
+                logger.error(f"对话CLI调用异常: {exc}")
+                return None
+
+        return await asyncio.to_thread(_run)
+
+    def _get_history(self, user_id: str) -> Deque[Tuple[str, str]]:
+        """获取或初始化用户对话历史"""
+        history = self.general_histories.get(user_id)
+        if history is None:
+            history = deque(maxlen=12)
+            self.general_histories[user_id] = history
+        return history
+
+    def _remember_exchange(self, user_id: str, user_text: str, assistant_text: str):
+        """记录一次对话轮次"""
+        history = self._get_history(user_id)
+        history.append(("user", user_text.strip()))
+        history.append(("assistant", assistant_text.strip()))
+
+    def _sanitize_cli_output(self, output: str) -> Optional[str]:
+        """清理CLI输出，去除控制字符和冗余前缀"""
+        if not output:
+            return None
+
+        cleaned = output.replace("\r\n", "\n").replace("\r", "\n").strip()
+        cleaned = re.sub(r"\x1b\[[0-9;]*[A-Za-z]", "", cleaned)
+
+        lines = [line.strip() for line in cleaned.split("\n") if line.strip()]
+        if not lines:
+            return None
+
+        # 删除通用前缀
+        first_line = lines[0]
+        first_line = re.sub(r"^(assistant|codex|claude|response)[:：]\s*", "", first_line, flags=re.IGNORECASE)
+        lines[0] = first_line
+
+        cleaned_text = "\n".join(lines).strip()
+        return cleaned_text[:1200]
+
+# 全局实例
+finance_agent = FinanceAgentUltimate()
+
+async def handle_agent_query(query: str, context: Dict = None) -> str:
+    """主入口"""
+    if not hasattr(finance_agent, 'initialized'):
+        await finance_agent.initialize()
+        finance_agent.initialized = True
+    
+    return await finance_agent.process_command(query, context)