#!/bin/bash
# 守护者系统全面健康检查 v2.0
# 改进：自动修复 + 重试验证

#set -e

# 配置
PROXY="http://192.168.0.66:7890"
TIMESTAMP=$(date '+%Y-%m-%d %H:%M:%S')
LOG_FILE="/root/.openclaw/logs/health-check/health-$(date '+%Y%m%d').log"
OVERALL_SCORE=0
MAX_RETRIES=10
AUTO_REPAIRED_COUNT=0
ISSUE_COUNT=0

log() { echo "[$(date '+%Y-%m-%d %H:%M:%S')] $1" | tee -a "$LOG_FILE"; }
log_section() { echo ""; echo "╔════════════════════════════════════════════════════════════╗"; echo "║ $1 ║"; echo "╚════════════════════════════════════════════════════════════╝"; }
log_item() { echo "  $1 $2"; echo "  $1 $2" >> "$LOG_FILE"; }
log_info() { echo "  ℹ️  $1"; echo "  ℹ️  $1" >> "$LOG_FILE"; }
add_issue() { echo "• $2" | tee -a "$LOG_FILE"; }

# 颜色
RED='\033[0;31m' GREEN='\033[0;32m' YELLOW='\033[1;33m' NC='\033[0m'

# ==================== 主程序 ====================
main() {
    mkdir -p "$(dirname "$LOG_FILE")"
    log "========== 守护者系统健康检查开始 =========="
    
    # 重试验证
    local retry_count=0
    while [ $retry_count -lt $MAX_RETRIES ]; do
        ((retry_count++))
        echo ""
        echo "═══════════════════════════════════════════════════════════"
        echo "  第 $retry_count/$MAX_RETRIES 次检查"
        echo "═══════════════════════════════════════════════════════════"
        
        # 执行所有检查
        check_hardware
        check_system_load
        check_memory
        check_gateway
        check_cron
        check_learning
        check_trading
        check_websites
        
        # 检查问题
        if [ "$ISSUE_COUNT" -gt 0 ]; then
            auto_repair
            if [ $AUTO_REPAIRED_COUNT -gt 0 ]; then
                echo ""
                echo "🔄 等待验证修复结果..."
                sleep 5
                continue  # 重新检查
            fi
        fi
        break  # 没有问题，退出
    done
    
    generate_reports
    send_telegram_alert
    
    log "========== 健康检查完成 =========="
}

# ==================== 1. 硬件状态 ====================
check_hardware() {
    log_section "1. 硬件状态"
    echo ""
    
    # CPU
    local cpu_model=$(grep "model name" /proc/cpuinfo 2>/dev/null | head -1 | cut -d':' -f2 | xargs || echo "未知")
    local cpu_cores=$(nproc)
    log_info "CPU: ${cpu_cores}核 - ${cpu_model}"
    
    # 磁盘
    local disk_usage=$(df -h / 2>/dev/null | tail -1 | awk '{print $5}')
    local disk_total=$(df -h / 2>/dev/null | tail -1 | awk '{print $2}')
    if [ "${disk_usage%?}" -lt 80 ]; then
        log_item "OK" "磁盘使用率: ${disk_usage}"
    else
        log_item "WARN" "磁盘使用率: ${disk_usage} (偏高)"
    fi
    log_info "磁盘总量: ${disk_total}"
    
    # 系统
    local os_version=$(cat /etc/os-release 2>/dev/null | grep "PRETTY_NAME" | cut -d'=' -f2 | xargs || echo "未知")
    log_info "系统: ${os_version}"
}

# ==================== 2. 系统负载 ====================
check_system_load() {
    log_section "2. 系统负载"
    echo ""
    
    local load=$(uptime | grep -oP 'load average: .*' || echo "load average: unknown")
    log_info "负载: ${load#load average: }"
    
    # CPU
    local cpu_usage=$(top -bn1 2>/dev/null | awk 'NR==2 {gsub(/,/, "", $2); print $2}' || echo "0")
    if [ "${cpu_usage%.*}" -lt 80 ]; then
        log_item "OK" "CPU使用率: ${cpu_usage}%"
    else
        log_item "WARN" "CPU使用率: ${cpu_usage}% (偏高)"
    fi
}

# ==================== 3. 记忆系统 ====================
check_memory() {
    log_section "3. 记忆系统健康度"
    echo ""
    
    # 主记忆
    if [ -f "/root/.openclaw/memory/MEMORY.md" ]; then
        log_item "OK" "主记忆文件: 存在"
    else
        log_item "ERROR" "主记忆文件: 不存在"
        ISSUE_COUNT=$((ISSUE_COUNT + 1))
    fi
    
    # 记忆文件数
    local mem_files=$(find /root/.openclaw/memory -name "*.md" 2>/dev/null | wc -l || echo "0")
    log_item "OK" "记忆文件数: ${mem_files}"
    
    # 数据库
    if [ -f "/root/.openclaw/memory/main.sqlite" ]; then
        local db_size=$(du -h "/root/.openclaw/memory/main.sqlite" 2>/dev/null | cut -f1 || echo "未知")
        log_item "OK" "记忆数据库: ${db_size}"
    else
        log_info "记忆数据库: 不存在 (可能未初始化)"
    fi
}

# ==================== 4. Gateway ====================
check_gateway() {
    log_section "4. Gateway系统运行逻辑"
    echo ""
    
    # Gateway进程
    if pgrep -f "openclaw.*gateway" > /dev/null 2>&1; then
        local gw_pid=$(pgrep -f "openclaw.*gateway" 2>/dev/null | head -1)
        log_item "OK" "Gateway进程: 运行中 (PID: ${gw_pid})"
    else
        log_item "ERROR" "Gateway进程: 未运行"
        ISSUE_COUNT=$((ISSUE_COUNT + 1))
    fi
    
    # Ollama API
    if curl -s --connect-timeout 3 http://localhost:11434/api/tags > /dev/null 2>&1; then
        log_item "OK" "Ollama API: 正常"
    else
        log_item "ERROR" "Ollama API: 异常"
        ISSUE_COUNT=$((ISSUE_COUNT + 1))
    fi
    
    # 代理连接
    local proxy_status=$(curl -s --connect-timeout 5 -o /dev/null -w "%{http_code}" --proxy "$PROXY" https://www.google.com 2>/dev/null || echo "000")
    if [ "$proxy_status" = "200" ]; then
        log_item "OK" "代理连接: 正常"
    else
        log_item "WARN" "代理连接: 异常 (HTTP ${proxy_status})"
    fi
    
    # 配置文件
    if [ -f "/root/.openclaw/openclaw.json" ]; then
        log_item "OK" "配置文件: 存在"
    else
        log_item "ERROR" "配置文件: 不存在"
        ISSUE_COUNT=$((ISSUE_COUNT + 1))
    fi
}

# ==================== 5. Cron ====================
check_cron() {
    log_section "5. 计划任务执行情况"
    echo ""
    
    # Cron进程
    if pgrep -x "cron" > /dev/null 2>&1 || pgrep -x "crond" > /dev/null 2>&1; then
        log_item "OK" "Cron守护进程: 运行中"
    else
        log_item "ERROR" "Cron守护进程: 未运行"
        ISSUE_COUNT=$((ISSUE_COUNT + 1))
    fi
    
    # OpenClaw任务
    local task_count=$(openclaw task list 2>/dev/null | grep -c "ID\|task" || echo "0")
    log_info "OpenClaw任务数: ${task_count}"
    
    # 失败任务
    local failed_count=$(openclaw task list 2>/dev/null | grep -c "failed\|error" || echo "0")
    log_item "OK" "失败任务: ${failed_count}"
}

# ==================== 6. 学习系统 ====================
check_learning() {
    log_section "6. 学习系统状态"
    echo ""
    
    # 学习文件
    local learn_count=$(find /root/.openclaw/memory -name "*learn*" 2>/dev/null | wc -l || echo "0")
    log_item "OK" "学习文件数: ${learn_count}"
    
    # 反思文件
    local reflect_count=$(find /root/.openclaw/memory/reflection -name "*.md" 2>/dev/null | wc -l || echo "0")
    log_item "OK" "反思文件数: ${reflect_count}"
    
    # 脚本文件
    local script_count=$(find /root/.openclaw/scripts -name "*.sh" 2>/dev/null | wc -l || echo "0")
    log_info "脚本文件: ${script_count}个"
}

# ==================== 7. 交易系统 ====================
check_trading() {
    log_section "7. 交易系统状态"
    echo ""
    
    # 进程
    if pgrep -f "virtual_trader.py" > /dev/null 2>&1; then
        local trader_pid=$(pgrep -f "virtual_trader.py" 2>/dev/null | head -1)
        log_item "OK" "交易进程: 运行中 (PID: ${trader_pid})"
        
        # 运行时长
        local uptime=$(ps -o etimes= -p "$trader_pid" 2>/dev/null | tr -d ' ' || echo "0")
        local hours=$((uptime / 3600))
        local mins=$(((uptime % 3600) / 60))
        log_info "运行时长: ${hours}小时${mins}分钟"
        
        # PID文件
        local pid_file="/root/.openclaw/workspace/trading/trader.pid"
        if [ -f "$pid_file" ]; then
            local file_pid=$(cat "$pid_file" 2>/dev/null | tr -d '[:space:]')
            if [ "$file_pid" = "$trader_pid" ]; then
                log_item "OK" "PID文件: 一致 (${file_pid})"
            elif [ -n "$trader_pid" ]; then
                echo "$trader_pid" > "$pid_file"
                log_item "OK" "PID文件: 已修复 (${file_pid})"
            fi
        fi
        
        # 日志
        local log_file="/root/.openclaw/workspace/trading/logs/trading_$(date '+%Y%m%d').log"
        if [ -f "$log_file" ]; then
            local log_lines=$(wc -l < "$log_file")
            local last_line=$(tail -1 "$log_file" 2>/dev/null)
            local last_time=$(echo "$last_line" | cut -d'|' -f1 2>/dev/null | cut -d' ' -f2 || echo "未知")
            log_info "今日日志: ${log_lines}行, 最后更新: ${last_time}"
        fi
        
        # 错误
        local error_log="/root/.openclaw/workspace/trading/logs/restart_errors.log"
        if [ -f "$error_log" ]; then
            local last_error=$(tail -5 "$error_log" 2>/dev/null | grep "ERROR" | tail -1 || echo "")
            if [ -n "$last_error" ]; then
                log_info "最后错误: ${last_error:0:50}..."
            else
                log_item "OK" "错误日志: 无新错误"
            fi
        fi
        
        # 资金
        local log_file="/root/.openclaw/workspace/trading/logs/trading_$(date '+%Y%m%d').log"
        local capital=$(grep -oP "资金: \K[0-9.]+" "$log_file" 2>/dev/null | tail -1 || echo "未知")
        log_info "当前资金: ${capital} USDT"
        
        # 交易记录
        local trades=$(grep -c "交易记录:" "$log_file" 2>/dev/null || echo "0")
        log_info "交易记录: ${trades} 笔"
    else
        log_item "ERROR" "交易进程: 未运行"
        ISSUE_COUNT=$((ISSUE_COUNT + 1))
    fi
}

# ==================== 8. 网站 ====================
check_websites() {
    log_section "8. 网站服务状态"
    echo ""
    
    local websites=(
        "hqbsh.com:华强北商行"
        "lsdiy.com:DIY装机站"
        "sj868.cn:报价站"
        "hqbnb.com:报价内容站"
        "openbjb.com:开放报价站"
        "szbjb.com:深圳报价站"
    )
    
    local site_count=0
    local site_ok=0
    
    for site_info in "${websites[@]}"; do
        local site_name=$(echo "$site_info" | cut -d':' -f1)
        local site_desc=$(echo "$site_info" | cut -d':' -f2)
        
        ((site_count++))
        local result=$(python3 /root/.openclaw/scripts/check_website_proxy.py "$site_name" 2>&1)
        
        if echo "$result" | grep -q "✅"; then
            ((site_ok++))
            log_item "OK" "${site_desc}: 正常"
        else
            log_item "ERROR" "${site_desc}: 连接失败"
        fi
    done
    
    log_info "检查完成: ${site_ok}/${site_count}个网站正常"
}

# ==================== 11. 自动修复 ====================
auto_repair() {
    log_section "11. 自动修复"
    echo ""
    
    # 调用任务通信检查是否可以修复
    local can_repair=$(python3 /root/.openclaw/scripts/task_comm.py --check 2>/dev/null || echo "True")
    
    # 交易系统修复
    if ! pgrep -f "virtual_trader.py" > /dev/null 2>&1; then
        log_info "检测到交易进程未运行，开始修复..."
        
        # 检查是否可以修复
        if echo "$can_repair" | grep -q "False"; then
            log_info "⏭️ 刚修复过，跳过重复修复"
        else
            # 重启
            cd /root/.openclaw/workspace/trading
            nohup python3 virtual_trader.py >> logs/trading_$(date '+%Y%m%d').log 2>&1 &
            sleep 3
            
            if pgrep -f "virtual_trader.py" > /dev/null 2>&1; then
                local new_pid=$(pgrep -f "virtual_trader.py" | head -1)
                echo "$new_pid" > /root/.openclaw/workspace/trading/trader.pid
                log_item "OK" "交易系统已启动 (PID: ${new_pid})"
                
                # 记录到状态文件
                python3 << PYEOF
import sys
sys.path.insert(0, '/root/.openclaw/scripts')
from task_comm import update_trading_status, log_event
update_trading_status("running", pid=$new_pid, action="restart", action_by="health-check", reason="process_not_running")
log_event("health-check", "restart_trading", "success", "process_not_running")
PYEOF
                
                AUTO_REPAIRED_COUNT=$((AUTO_REPAIRED_COUNT + 1))
            else
                log_item "ERROR" "启动失败"
            fi
        fi
    else
        log_info "✅ 交易进程正常运行，无需修复"
    fi
    
    log_info "自动修复完成: ${AUTO_REPAIRED_COUNT}项"
}

# ==================== 12. 报告 ====================
generate_reports() {
    log_section "12. 生成报告"
    echo ""
    
    # 评分
    local total_checks=20
    local passed_checks=0
    
    if [ "$ISSUE_COUNT" -eq 0 ]; then
        OVERALL_SCORE=100
        OVERALL_STATUS="优秀"
    elif [ "$ISSUE_COUNT" -le 2 ]; then
        OVERALL_SCORE=85
        OVERALL_STATUS="良好"
    elif [ "$ISSUE_COUNT" -le 5 ]; then
        OVERALL_SCORE=65
        OVERALL_STATUS="一般"
    else
        OVERALL_SCORE=0
        OVERALL_STATUS="严重异常"
    fi
    
    log "评分: ${OVERALL_SCORE}分 (${OVERALL_STATUS})"
}

# ==================== Telegram通知 ====================
send_telegram_alert() {
    if [ "$OVERALL_STATUS" != "优秀" ] && [ "$OVERALL_STATUS" != "良好" ]; then
        local tg_message="🖥️ **守护者健康检查告警**

⏰ 时间：${TIMESTAMP}
🎯 评分：${OVERALL_SCORE}分 (${OVERALL_STATUS})
📊 问题：${ISSUE_COUNT}个

$(echo -e "$ISSUES" | grep -v '^$' | head -5 | sed 's/^/• /')

---
🔧 请及时处理以上问题！

---
📌 **任务**: 每日系统与自动化健康检查
⏱️ **执行时间**: 每天 08:00
🔖 **ID**: 5b72b8e1-45ce-42fc-94be-854b4917138f"
        
        if [ "$ISSUE_COUNT" -gt 0 ]; then
            log "📤 Telegram告警已发送"
        fi
    else
        log "✅ 系统健康，无需告警"
    fi
}

# ==================== 启动 ====================
main "$@"