commit;

README.md

@@ -211,6 +211,127 @@ pip install -r requirements.txt
 
 筛选逻辑：选择行业环境稳定、有高质量新业务合作、财务状况良好、券商看好且上涨空间大的企业。
 
+## Docker 部署说明
+
+本系统支持通过 Docker 进行部署，可以单实例部署或多实例部署。
+
+### 1. 单实例部署
+
+```bash
+# 构建镜像
+docker-compose build
+
+# 启动服务
+docker-compose up -d
+```
+
+### 2. 多实例部署
+
+提供了三个脚本用于管理多实例部署，每个实例将在不同端口上运行（从5088开始递增）：
+
+#### 使用 docker run 部署多实例（推荐）
+
+```bash
+# 赋予脚本执行权限
+chmod +x deploy-multiple.sh
+
+# 部署3个实例，端口将为5088、5089、5090
+./deploy-multiple.sh 3
+```
+
+每个实例将有自己独立的数据目录：`./instances/instance-N/`，包含独立的日志和报告文件。
+
+#### 使用 docker-compose 部署多实例
+
+```bash
+# 赋予脚本执行权限
+chmod +x deploy-compose.sh
+
+# 部署3个实例，端口将为5088、5089、5090
+./deploy-compose.sh 3
+```
+
+每个实例将使用独立的项目名称 `stock-app-N`。
+
+#### 管理多实例
+
+使用 `manage-instances.sh` 脚本可以方便地管理已部署的实例：
+
+```bash
+# 赋予脚本执行权限
+chmod +x manage-instances.sh
+
+# 查看所有实例状态
+./manage-instances.sh status
+
+# 列出正在运行的实例
+./manage-instances.sh list
+
+# 启动指定实例
+./manage-instances.sh start 2
+
+# 启动所有实例
+./manage-instances.sh start all
+
+# 停止指定实例
+./manage-instances.sh stop 1
+
+# 停止所有实例
+./manage-instances.sh stop all
+
+# 重启指定实例
+./manage-instances.sh restart 3
+
+# 查看指定实例的日志
+./manage-instances.sh logs 1
+
+# 删除指定实例
+./manage-instances.sh remove 2
+
+# 删除所有实例
+./manage-instances.sh remove all
+```
+
+### 3. 中文字体安装
+
+系统需要中文字体以正常生成PDF报告。如果部署时遇到 `找不到中文字体文件` 错误，可以使用以下方法解决：
+
+#### 自动安装字体（推荐）
+
+```bash
+# 进入容器内部
+docker exec -it [容器ID或名称] bash
+
+# 在容器内执行字体安装脚本
+cd /app
+python src/fundamentals_llm/setup_fonts.py
+```
+
+#### 手动安装字体
+
+1. 在宿主机上下载中文字体文件（如simhei.ttf、wqy-microhei.ttc等）
+2. 创建字体目录并复制字体文件：
+```bash
+mkdir -p src/fundamentals_llm/fonts
+cp 你下载的字体文件.ttf src/fundamentals_llm/fonts/simhei.ttf
+```
+3. 重新构建镜像：
+```bash
+docker-compose build
+```
+
+### 4. 访问服务
+
+部署完成后，可以通过以下URL访问服务：
+
+- 单实例：`http://服务器IP:5000`
+- 多实例：`http://服务器IP:508X`（X为实例序号，从8开始）
+
+例如：
+- 实例1：`http://服务器IP:5088`
+- 实例2：`http://服务器IP:5089`
+- 实例3：`http://服务器IP:5090`
+
 ## 数据库查询示例
 
 以下SQL查询示例展示了如何从数据库中筛选特定类型的企业：
@@ -251,3 +372,11 @@ AND stock_code IN (
 ```
 
 注意：实际查询时可能需要根据数据库表结构调整SQL语句。建议根据具体需求组合使用不同的筛选条件。 
+
+
+
+# 重新构建镜像
+docker-compose build
+
+# 重启所有实例
+./manage-instances.sh restart all

requirements.txt

@@ -13,5 +13,5 @@ volcengine-python-sdk[ark]
 openai>=1.0
 reportlab>=4.3.1
 markdown2>=2.5.3
-# 1. 按下 Win+R ，输入 regedit 打开注册表编辑器。
-# 2. 设置 \HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\FileSystem 路径下的变量 LongPathsEnabled 为 1 即可。
+google-genai
+redis==5.2.1

src/app.py

@@ -1,27 +1,47 @@
 import sys
 import os
+from datetime import datetime, timedelta
+import pandas as pd
+import uuid
+import json
+from threading import Thread
 
 from src.fundamentals_llm.fundamental_analysis_database import get_analysis_result, get_db
 
 # 添加项目根目录到 Python 路径
 sys.path.append(os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
 
-from flask import Flask, jsonify, request
+from flask import Flask, jsonify, request, send_from_directory
 from flask_cors import CORS
 import logging
 
 # 导入企业筛选器
 from src.fundamentals_llm.enterprise_screener import EnterpriseScreener
 
+# 导入股票回测器
+from src.stock_analysis_v2 import run_backtest, StockBacktester
+
 # 设置日志
 logging.basicConfig(
     level=logging.INFO,
-    format='%(asctime)s - %(name)s - %(levelname)s - %(message)s'
+    format='%(asctime)s - %(name)s - %(levelname)s - %(message)s',
+    handlers=[
+        logging.StreamHandler(),  # 输出到控制台
+        logging.FileHandler(f'logs/app_{datetime.now().strftime("%Y%m%d")}.log', encoding='utf-8')  # 输出到文件
+    ]
 )
+
+# 确保logs和results目录存在
+os.makedirs('logs', exist_ok=True)
+os.makedirs('results', exist_ok=True)
+os.makedirs('results/tasks', exist_ok=True)
+os.makedirs('static/results', exist_ok=True)
+
 logger = logging.getLogger(__name__)
+logger.info("Flask应用启动")
 
 # 创建 Flask 应用
-app = Flask(__name__)
+app = Flask(__name__, static_folder='static')
 CORS(app)  # 启用跨域请求支持
 
 # 创建企业筛选器实例
@@ -30,6 +50,329 @@ screener = EnterpriseScreener()
 # 获取数据库连接
 db = next(get_db())
 
+# 获取项目根目录
+ROOT_DIR = os.path.dirname(os.path.dirname(os.path.abspath(__file__)))
+REPORTS_DIR = os.path.join(ROOT_DIR, 'src', 'reports')
+
+# 确保reports目录存在
+os.makedirs(REPORTS_DIR, exist_ok=True)
+logger.info(f"报告目录路径: {REPORTS_DIR}")
+
+# 存储回测任务状态的字典
+backtest_tasks = {}
+
+def run_backtest_task(task_id, stocks_buy_dates, end_date):
+    """
+    在后台运行回测任务
+    """
+    try:
+        logger.info(f"开始执行回测任务 {task_id}")
+        # 更新任务状态为进行中
+        backtest_tasks[task_id]['status'] = 'running'
+        
+        # 运行回测
+        results, stats_list = run_backtest(stocks_buy_dates, end_date)
+        
+        # 如果回测成功
+        if results and stats_list:
+            # 计算总体统计
+            stats_df = pd.DataFrame(stats_list)
+            total_profit = stats_df['final_profit'].sum()
+            avg_win_rate = stats_df['win_rate'].mean()
+            avg_holding_days = stats_df['avg_holding_days'].mean()
+            
+            # 找出最佳止盈比例(假设为0.15，实际应从回测结果中分析得出)
+            best_take_profit_pct = 0.15
+            
+            # 获取图表URL
+            chart_urls = {
+                "all_stocks": f"/static/results/{task_id}/all_stocks_analysis.png",
+                "profit_matrix": f"/static/results/{task_id}/profit_matrix_analysis.png"
+            }
+            
+            # 保存股票详细统计
+            stock_stats = []
+            for stat in stats_list:
+                stock_stats.append({
+                    "symbol": stat['symbol'],
+                    "total_trades": int(stat['total_trades']),
+                    "profitable_trades": int(stat['profitable_trades']),
+                    "loss_trades": int(stat['loss_trades']),
+                    "win_rate": float(stat['win_rate']),
+                    "avg_holding_days": float(stat['avg_holding_days']),
+                    "final_profit": float(stat['final_profit']),
+                    "entry_count": int(stat['entry_count'])
+                })
+            
+            # 构建结果数据
+            result_data = {
+                "task_id": task_id,
+                "status": "completed",
+                "results": {
+                    "total_profit": float(total_profit),
+                    "win_rate": float(avg_win_rate),
+                    "avg_holding_days": float(avg_holding_days),
+                    "best_take_profit_pct": best_take_profit_pct,
+                    "stock_stats": stock_stats,
+                    "chart_urls": chart_urls
+                }
+            }
+            
+            # 保存结果到文件
+            task_result_path = os.path.join('results', 'tasks', f"{task_id}.json")
+            with open(task_result_path, 'w', encoding='utf-8') as f:
+                json.dump(result_data, f, ensure_ascii=False, indent=2)
+            
+            # 更新任务状态为已完成
+            backtest_tasks[task_id].update({
+                'status': 'completed',
+                'results': result_data['results']
+            })
+            
+            logger.info(f"回测任务 {task_id} 已完成")
+        else:
+            # 更新任务状态为失败
+            backtest_tasks[task_id]['status'] = 'failed'
+            backtest_tasks[task_id]['error'] = "回测未产生有效结果"
+            logger.error(f"回测任务 {task_id} 失败：回测未产生有效结果")
+    except Exception as e:
+        # 更新任务状态为失败
+        backtest_tasks[task_id]['status'] = 'failed'
+        backtest_tasks[task_id]['error'] = str(e)
+        logger.error(f"回测任务 {task_id} 失败：{str(e)}")
+
+@app.route('/api/backtest/run', methods=['POST'])
+def start_backtest():
+    """启动回测任务
+    
+    请求体格式:
+    {
+        "stocks_buy_dates": {
+            "SH600522": ["2022-05-10", "2022-06-10"],  // 股票代码: [买入日期列表]
+            "SZ002340": ["2022-06-15"],
+            "SH601615": ["2022-07-20", "2022-08-01"]
+        },
+        "end_date": "2022-10-20"  // 所有股票共同的结束日期
+    }
+    
+    返回内容:
+    {
+        "task_id": "backtask-khkerhy4u237y489237489truiy8432"
+    }
+    """
+    try:
+        # 从请求体获取参数
+        data = request.get_json()
+        
+        if not data:
+            return jsonify({
+                "status": "error", 
+                "message": "请求格式错误: 需要提供JSON数据"
+            }), 400
+            
+        stocks_buy_dates = data.get('stocks_buy_dates')
+        end_date = data.get('end_date')
+        
+        if not stocks_buy_dates or not isinstance(stocks_buy_dates, dict):
+            return jsonify({
+                "status": "error", 
+                "message": "请求格式错误: 需要提供stocks_buy_dates字典"
+            }), 400
+            
+        if not end_date or not isinstance(end_date, str):
+            return jsonify({
+                "status": "error", 
+                "message": "请求格式错误: 需要提供有效的end_date"
+            }), 400
+            
+        # 验证日期格式
+        try:
+            datetime.strptime(end_date, '%Y-%m-%d')
+            for stock_code, buy_dates in stocks_buy_dates.items():
+                if not isinstance(buy_dates, list):
+                    return jsonify({
+                        "status": "error", 
+                        "message": f"请求格式错误: 股票 {stock_code} 的买入日期必须是列表"
+                    }), 400
+                for buy_date in buy_dates:
+                    datetime.strptime(buy_date, '%Y-%m-%d')
+        except ValueError as e:
+            return jsonify({
+                "status": "error", 
+                "message": f"日期格式错误: {str(e)}"
+            }), 400
+        
+        # 生成任务ID
+        task_id = f"backtask-{uuid.uuid4().hex[:16]}"
+        
+        # 创建任务目录
+        task_dir = os.path.join('static', 'results', task_id)
+        os.makedirs(task_dir, exist_ok=True)
+        
+        # 记录任务信息
+        backtest_tasks[task_id] = {
+            'status': 'pending', 
+            'created_at': datetime.now().isoformat(),
+            'stocks_buy_dates': stocks_buy_dates,
+            'end_date': end_date
+        }
+        
+        # 创建线程运行回测
+        thread = Thread(target=run_backtest_task, args=(task_id, stocks_buy_dates, end_date))
+        thread.daemon = True
+        thread.start()
+        
+        logger.info(f"已创建回测任务: {task_id}")
+        
+        return jsonify({
+            "task_id": task_id
+        })
+        
+    except Exception as e:
+        logger.error(f"创建回测任务失败: {str(e)}")
+        return jsonify({
+            "status": "error", 
+            "message": f"创建回测任务失败: {str(e)}"
+        }), 500
+
+@app.route('/api/backtest/status', methods=['GET'])
+def check_backtest_status():
+    """查询回测任务状态
+    
+    参数:
+    - task_id: 回测任务ID
+    
+    返回内容:
+    {
+        "task_id": "backtask-khkerhy4u237y489237489truiy8432",
+        "status": "running" | "completed" | "failed",
+        "created_at": "2023-12-01T10:30:45",
+        "error": "错误信息（如有）"
+    }
+    """
+    try:
+        task_id = request.args.get('task_id')
+        
+        if not task_id:
+            return jsonify({
+                "status": "error", 
+                "message": "请求格式错误: 需要提供task_id参数"
+            }), 400
+        
+        # 检查任务是否存在
+        if task_id not in backtest_tasks:
+            return jsonify({
+                "status": "error", 
+                "message": f"任务不存在: {task_id}"
+            }), 404
+        
+        # 获取任务信息
+        task_info = backtest_tasks[task_id]
+        
+        # 构建响应
+        response = {
+            "task_id": task_id,
+            "status": task_info['status'],
+            "created_at": task_info['created_at']
+        }
+        
+        # 如果任务失败，添加错误信息
+        if task_info['status'] == 'failed' and 'error' in task_info:
+            response['error'] = task_info['error']
+        
+        return jsonify(response)
+        
+    except Exception as e:
+        logger.error(f"查询任务状态失败: {str(e)}")
+        return jsonify({
+            "status": "error", 
+            "message": f"查询任务状态失败: {str(e)}"
+        }), 500
+
+@app.route('/api/backtest/result', methods=['GET'])
+def get_backtest_result():
+    """获取回测任务结果
+    
+    参数:
+    - task_id: 回测任务ID
+    
+    返回内容:
+    {
+        "task_id": "backtask-2023121-001",
+        "status": "completed",
+        "results": {
+            "total_profit": 123456.78,
+            "win_rate": 75.5,
+            "avg_holding_days": 12.3,
+            "best_take_profit_pct": 0.15,
+            "stock_stats": [...],
+            "chart_urls": {
+                "all_stocks": "/static/results/backtask-2023121-001/all_stocks_analysis.png",
+                "profit_matrix": "/static/results/backtask-2023121-001/profit_matrix_analysis.png"
+            }
+        }
+    }
+    """
+    try:
+        task_id = request.args.get('task_id')
+        
+        if not task_id:
+            return jsonify({
+                "status": "error", 
+                "message": "请求格式错误: 需要提供task_id参数"
+            }), 400
+        
+        # 检查任务是否存在
+        if task_id not in backtest_tasks:
+            # 尝试从文件加载
+            task_result_path = os.path.join('results', 'tasks', f"{task_id}.json")
+            if os.path.exists(task_result_path):
+                with open(task_result_path, 'r', encoding='utf-8') as f:
+                    result_data = json.load(f)
+                return jsonify(result_data)
+            else:
+                return jsonify({
+                    "status": "error", 
+                    "message": f"任务不存在: {task_id}"
+                }), 404
+        
+        # 获取任务信息
+        task_info = backtest_tasks[task_id]
+        
+        # 检查任务是否完成
+        if task_info['status'] != 'completed':
+            return jsonify({
+                "status": "error", 
+                "message": f"任务尚未完成或已失败: {task_info['status']}"
+            }), 400
+        
+        # 构建响应
+        response = {
+            "task_id": task_id,
+            "status": "completed",
+            "results": task_info['results']
+        }
+        
+        return jsonify(response)
+        
+    except Exception as e:
+        logger.error(f"获取任务结果失败: {str(e)}")
+        return jsonify({
+            "status": "error", 
+            "message": f"获取任务结果失败: {str(e)}"
+        }), 500
+
+@app.route('/api/reports/<path:filename>')
+def serve_report(filename):
+    """提供PDF报告的访问"""
+    try:
+        logger.info(f"请求文件: {filename}")
+        logger.info(f"从目录: {REPORTS_DIR} 提供文件")
+        return send_from_directory(REPORTS_DIR, filename, as_attachment=True)
+    except Exception as e:
+        logger.error(f"提供报告文件失败: {str(e)}")
+        return jsonify({"error": "文件不存在"}), 404
+
 @app.route('/api/health', methods=['GET'])
 def health_check():
     """健康检查接口"""
@@ -185,10 +528,10 @@ def generate_reports():
         
         # 导入 PDF 生成器模块
         try:
-            from src.fundamentals_llm.pdf_generator import generate_investment_report
+            from src.fundamentals_llm.pdf_generator import PDFGenerator
         except ImportError:
             try:
-                from fundamentals_llm.pdf_generator import generate_investment_report
+                from fundamentals_llm.pdf_generator import PDFGenerator
             except ImportError as e:
                 logger.error(f"无法导入 PDF 生成器模块: {str(e)}")
                 return jsonify({
@@ -200,8 +543,14 @@ def generate_reports():
         generated_reports = []
         for stock_code, stock_name in stocks:
             try:
+                # 创建 PDF 生成器实例
+                generator = PDFGenerator()
                 # 调用 PDF 生成器
-                report_path = generate_investment_report(stock_code, stock_name)
+                report_path = generator.generate_pdf(
+                    title=f"{stock_name}({stock_code}) 基本面分析报告",
+                    content_dict={},  # 这里需要传入实际的内容字典
+                    filename=f"{stock_name}_{stock_code}_analysis.pdf"
+                )
                 generated_reports.append({
                     "code": stock_code,
                     "name": stock_name,
@@ -505,9 +854,10 @@ def analyze_and_recommend():
             "message": f"分析和推荐股票失败: {str(e)}"
         }), 500
 
+
 @app.route('/api/comprehensive_analysis', methods=['POST'])
 def comprehensive_analysis():
-    """综合分析接口 - 组合多种功能和参数
+    """综合分析接口 - 使用队列方式处理被锁定的股票
     
     请求体格式:
     {
@@ -558,36 +908,9 @@ def comprehensive_analysis():
         if not isinstance(limit, int) or limit <= 0:
             limit = 10
         
-        # 导入必要的聊天机器人模块
+        # 导入必要的模块
         try:
-            # 首先尝试导入聊天机器人模块
-            try:
-                from src.fundamentals_llm.chat_bot import ChatBot as OnlineChatBot
-                logger.info("成功从 src.fundamentals_llm.chat_bot 导入 ChatBot")
-            except ImportError as e1:
-                try:
-                    from fundamentals_llm.chat_bot import ChatBot as OnlineChatBot
-                    logger.info("成功从 fundamentals_llm.chat_bot 导入 ChatBot")
-                except ImportError as e2:
-                    logger.error(f"无法导入在线聊天机器人模块: {str(e1)}, {str(e2)}")
-                    return jsonify({
-                        "status": "error", 
-                        "message": f"服务器配置错误: 聊天机器人模块不可用，错误详情: {str(e2)}"
-                    }), 500
-                    
-            # 然后尝试导入离线聊天机器人模块
-            try:
-                from src.fundamentals_llm.chat_bot_with_offline import ChatBot as OfflineChatBot
-                logger.info("成功从 src.fundamentals_llm.chat_bot_with_offline 导入 ChatBot")
-            except ImportError as e1:
-                try:
-                    from fundamentals_llm.chat_bot_with_offline import ChatBot as OfflineChatBot
-                    logger.info("成功从 fundamentals_llm.chat_bot_with_offline 导入 ChatBot")
-                except ImportError as e2:
-                    logger.warning(f"无法导入离线聊天机器人模块: {str(e1)}, {str(e2)}")
-                    # 这里可以继续执行，因为某些功能可能不需要离线模型
-            
-            # 最后导入基本面分析器
+            # 先导入基本面分析器
             try:
                 from src.fundamentals_llm.fundamental_analysis import FundamentalAnalyzer
                 logger.info("成功从 src.fundamentals_llm.fundamental_analysis 导入 FundamentalAnalyzer")
@@ -601,6 +924,19 @@ def comprehensive_analysis():
                         "status": "error", 
                         "message": f"服务器配置错误: 基本面分析模块不可用，错误详情: {str(e2)}"
                     }), 500
+            
+            # 再导入其他可能需要的模块
+            try:
+                from src.fundamentals_llm.chat_bot import ChatBot as OnlineChatBot
+                from src.fundamentals_llm.chat_bot_with_offline import ChatBot as OfflineChatBot
+            except ImportError:
+                try:
+                    from fundamentals_llm.chat_bot import ChatBot as OnlineChatBot
+                    from fundamentals_llm.chat_bot_with_offline import ChatBot as OfflineChatBot
+                except ImportError:
+                    # 这些模块不是必须的，所以继续执行
+                    logger.warning("无法导入聊天机器人模块，但这不会影响基本功能")
+                    
         except Exception as e:
             logger.error(f"导入必要模块时出错: {str(e)}")
             return jsonify({
@@ -611,64 +947,141 @@ def comprehensive_analysis():
         # 创建基本面分析器实例
         analyzer = FundamentalAnalyzer()
         
-        # 为每个股票生成投资建议
-        investment_advices = []
-        for stock_code, stock_name in stocks:
+        # 准备结果容器
+        investment_advices = {}  # 使用字典，股票代码作为键
+        processing_queue = list(stocks)  # 初始处理队列
+        max_attempts = 5  # 最大重试次数
+        total_attempts = 0
+        
+        # 导入数据库模块
+        from src.fundamentals_llm.fundamental_analysis_database import get_analysis_result, get_db
+        
+        # 开始处理队列
+        while processing_queue and total_attempts < max_attempts:
+            total_attempts += 1
+            logger.info(f"开始第 {total_attempts} 轮处理，队列中有 {len(processing_queue)} 只股票")
+            
+            # 暂存下一轮需要处理的股票
+            next_round_queue = []
+            
+            # 处理当前队列中的所有股票
+            for stock_code, stock_name in processing_queue:
                 try:
-                # 生成投资建议
+                    # 检查是否已有分析结果
+                    db = next(get_db())
+                    existing_result = get_analysis_result(db, stock_code, "investment_advice")
+                    
+                    # 如果已有近期结果，直接使用
+                    if existing_result and existing_result.update_time > datetime.now() - timedelta(hours=12):
+                        investment_advices[stock_code] = {
+                            "code": stock_code,
+                            "name": stock_name,
+                            "advice": existing_result.ai_response,
+                            "reasoning": existing_result.reasoning_process,
+                            "references": existing_result.references,
+                            "status": "success",
+                            "from_cache": True
+                        }
+                        logger.info(f"使用缓存的 {stock_name}({stock_code}) 分析结果")
+                        continue
+                    
+                    # 检查是否被锁定
+                    if analyzer.is_stock_locked(stock_code, "investment_advice"):
+                        # 已被锁定，放到下一轮队列
+                        next_round_queue.append([stock_code, stock_name])
+                        # 记录状态
+                        if stock_code not in investment_advices:
+                            investment_advices[stock_code] = {
+                                "code": stock_code,
+                                "name": stock_name,
+                                "status": "pending",
+                                "message": f"股票 {stock_code} 正在被其他请求分析中，已加入等待队列"
+                            }
+                        logger.info(f"股票 {stock_name}({stock_code}) 已被锁定，放入下一轮队列")
+                        continue
+                    
+                    # 尝试锁定并分析
+                    analyzer.lock_stock(stock_code, "investment_advice")
+                    try:
+                        # 执行分析
                         success, advice, reasoning, references = analyzer.query_analysis(
                             stock_code, stock_name, "investment_advice"
                         )
                         
+                        # 记录结果
                         if success:
-                    investment_advices.append({
+                            investment_advices[stock_code] = {
                                 "code": stock_code,
                                 "name": stock_name,
                                 "advice": advice,
                                 "reasoning": reasoning,
                                 "references": references,
                                 "status": "success"
-                    })
-                    logger.info(f"成功生成 {stock_name}({stock_code}) 的投资建议")
+                            }
+                            logger.info(f"成功分析 {stock_name}({stock_code})")
                         else:
-                    investment_advices.append({
+                            investment_advices[stock_code] = {
                                 "code": stock_code,
                                 "name": stock_name,
                                 "status": "error",
-                        "error": advice
-                    })
-                    logger.error(f"生成 {stock_name}({stock_code}) 的投资建议失败: {advice}")
+                                "error": advice or "分析失败，无详细信息"
+                            }
+                            logger.error(f"分析 {stock_name}({stock_code}) 失败: {advice}")
+                    finally:
+                        # 确保释放锁
+                        analyzer.unlock_stock(stock_code, "investment_advice")
+                        
                 except Exception as e:
-                logger.error(f"处理 {stock_name}({stock_code}) 时出错: {str(e)}")
-                investment_advices.append({
+                    # 处理异常
+                    investment_advices[stock_code] = {
                         "code": stock_code,
                         "name": stock_name,
                         "status": "error",
                         "error": str(e)
+                    }
+                    logger.error(f"处理 {stock_name}({stock_code}) 时出错: {str(e)}")
+                    # 确保释放锁
+                    try:
+                        analyzer.unlock_stock(stock_code, "investment_advice")
+                    except:
+                        pass
+            
+            # 如果还有下一轮要处理的股票，等待一段时间后继续
+            if next_round_queue:
+                logger.info(f"本轮结束，还有 {len(next_round_queue)} 只股票等待下一轮处理")
+                # 等待30秒再处理下一轮
+                import time
+                time.sleep(30)
+                processing_queue = next_round_queue
+            else:
+                # 所有股票都已处理完，退出循环
+                logger.info("所有股票处理完毕")
+                processing_queue = []
+        
+        # 处理仍在队列中的股票（达到最大重试次数但仍未处理的）
+        for stock_code, stock_name in processing_queue:
+            if stock_code in investment_advices and investment_advices[stock_code]["status"] == "pending":
+                investment_advices[stock_code].update({
+                    "status": "timeout",
+                    "message": f"等待超时，股票 {stock_code} 可能正在被长时间分析"
                 })
+                logger.warning(f"股票 {stock_name}({stock_code}) 分析超时")
+        
+        # 将字典转换为列表
+        investment_advice_list = list(investment_advices.values())
         
         # 生成PDF报告（如果需要）
         pdf_results = []
         if generate_pdf:
             try:
-                # 导入PDF生成器模块
+                # 针对已成功分析的股票生成PDF
+                for stock_info in investment_advice_list:
+                    if stock_info["status"] == "success":
+                        stock_code = stock_info["code"]
+                        stock_name = stock_info["name"]
                         try:
-                    from src.fundamentals_llm.pdf_generator import generate_investment_report
-                except ImportError:
-                    try:
-                        from fundamentals_llm.pdf_generator import generate_investment_report
-                    except ImportError as e:
-                        logger.error(f"无法导入 PDF 生成器模块: {str(e)}")
-                        return jsonify({
-                            "status": "error", 
-                            "message": f"服务器配置错误: PDF 生成器模块不可用, 错误详情: {str(e)}"
-                        }), 500
-                
-                # 生成报告
-                for stock_code, stock_name in stocks:
-                    try:
-                        # 调用 PDF 生成器
-                        report_path = generate_investment_report(stock_code, stock_name)
+                            report_path = analyzer.generate_pdf_report(stock_code, stock_name)
+                            if report_path:
                                 pdf_results.append({
                                     "code": stock_code,
                                     "name": stock_name,
@@ -676,6 +1089,14 @@ def comprehensive_analysis():
                                     "status": "success"
                                 })
                                 logger.info(f"成功生成 {stock_name}({stock_code}) 的投资报告: {report_path}")
+                            else:
+                                pdf_results.append({
+                                    "code": stock_code,
+                                    "name": stock_name,
+                                    "status": "error",
+                                    "error": "生成报告失败"
+                                })
+                                logger.error(f"生成 {stock_name}({stock_code}) 的投资报告失败")
                         except Exception as e:
                             logger.error(f"生成 {stock_name}({stock_code}) 的投资报告失败: {str(e)}")
                             pdf_results.append({
@@ -784,11 +1205,24 @@ def comprehensive_analysis():
                 logger.error(f"应用企业画像筛选失败: {str(e)}")
                 filtered_stocks = []
         
+        # 统计各种状态的股票数量
+        success_count = sum(1 for item in investment_advice_list if item["status"] == "success")
+        pending_count = sum(1 for item in investment_advice_list if item["status"] == "pending")
+        timeout_count = sum(1 for item in investment_advice_list if item["status"] == "timeout")
+        error_count = sum(1 for item in investment_advice_list if item["status"] == "error")
+        
         # 返回结果
         response = {
-            "status": "success",
+            "status": "success" if success_count > 0 else "partial_success" if success_count + pending_count > 0 else "failed",
             "total_input_stocks": len(stocks),
-            "investment_advices": investment_advices
+            "stats": {
+                "success": success_count,
+                "pending": pending_count,
+                "timeout": timeout_count,
+                "error": error_count
+            },
+            "rounds_attempted": total_attempts,
+            "investment_advices": investment_advice_list
         }
         
         if profile_filter:

src/fundamentals_llm/chat_bot.py

@@ -150,11 +150,15 @@ class ChatBot:
             logger.error(f"格式化参考资料时出错: {str(e)}")
             return str(ref)
         
-    def chat(self, user_input: str) -> Dict[str, Any]:
+    def chat(self, user_input: str, temperature: float = 1.0, top_p: float = 0.7, max_tokens: int = 4096, frequency_penalty: float = 0.0) -> Dict[str, Any]:
         """与AI进行对话
         
         Args:
             user_input: 用户输入的问题
+            temperature: 控制输出的随机性，范围0-2，默认1.0
+            top_p: 控制输出的多样性，范围0-1，默认0.7
+            max_tokens: 控制输出的最大长度，默认4096
+            frequency_penalty: 控制重复惩罚，范围-2到2，默认0.0
             
         Returns:
             Dict[str, Any]: 包含以下字段的字典：
@@ -173,6 +177,10 @@ class ChatBot:
             stream = self.client.chat.completions.create(
                 model=self.model,
                 messages=self.conversation_history,
+                temperature=temperature,
+                top_p=top_p,
+                max_tokens=max_tokens,
+                frequency_penalty=frequency_penalty,
                 stream=True
             )
             

src/fundamentals_llm/chat_bot_with_gemini.py

@@ -0,0 +1,420 @@
+import logging
+import os
+import sys
+import json
+from typing import Dict, List, Optional, Any, Union
+from pydantic import BaseModel, Field
+from google import genai
+from google.genai.types import Tool, GenerateContentConfig, GoogleSearch
+
+# 设置日志记录
+logger = logging.getLogger(__name__)
+
+# 定义基础数据结构
+class TextAnalysisResult(BaseModel):
+    """文本分析结果，包含分析正文、推理过程和引用URL"""
+    analysis_text: str = Field(description="详细的分析文本")
+    reasoning_process: Optional[str] = Field(description="模型的推理过程", default=None)
+    references: Optional[List[str]] = Field(description="参考资料和引用URL列表", default=None)
+
+class NumericalAnalysisResult(BaseModel):
+    """数值分析结果，包含数值和分析描述"""
+    value: str = Field(description="评估值")
+    description: str = Field(description="评估描述")
+
+# 分析类型到模型类的映射
+ANALYSIS_TYPE_MODEL_MAP = {
+    # 文本分析类型
+    "text_analysis_result": TextAnalysisResult,
+    
+    # 数值分析类型
+    "numerical_analysis_result": NumericalAnalysisResult
+}
+
+class JsonOutputParser:
+    """解析JSON输出的工具类"""
+    def __init__(self, pydantic_object):
+        """初始化解析器
+        
+        Args:
+            pydantic_object: 需要解析成的Pydantic模型类
+        """
+        self.pydantic_object = pydantic_object
+
+    def get_format_instructions(self) -> str:
+        """获取格式化指令"""
+        schema = self.pydantic_object.schema()
+        schema_str = json.dumps(schema, ensure_ascii=False, indent=2)
+        
+        return f"""请将你的回答格式化为符合以下JSON结构的格式:
+{schema_str}
+
+你的回答应该只包含一个有效的JSON对象，而不包含其他内容。请不要包含任何解释或前缀，仅输出JSON本身。
+"""
+
+    def parse(self, text: str) -> Any:
+        """解析文本为对象"""
+        try:
+            # 尝试解析JSON
+            text = text.strip()
+            # 如果文本以```json开头且以```结尾，则提取中间部分
+            if text.startswith("```json") and text.endswith("```"):
+                text = text[7:-3].strip()
+            # 如果文本以```开头且以```结尾，则提取中间部分
+            elif text.startswith("```") and text.endswith("```"):
+                text = text[3:-3].strip()
+                
+            json_obj = json.loads(text)
+            return self.pydantic_object.parse_obj(json_obj)
+        except Exception as e:
+            raise ValueError(f"无法解析为JSON或无法匹配模式: {e}")
+
+# 获取项目根目录的绝对路径
+ROOT_DIR = os.path.dirname(os.path.dirname(os.path.abspath(__file__)))
+
+# 尝试导入配置
+try:
+    # 添加项目根目录到 Python 路径
+    sys.path.append(os.path.dirname(ROOT_DIR))
+    sys.path.append(ROOT_DIR)
+
+    # 导入配置
+    try:
+        from scripts.config import get_model_config
+
+        logger.info("成功从scripts.config导入配置")
+    except ImportError:
+        try:
+            from src.scripts.config import get_model_config
+
+            logger.info("成功从src.scripts.config导入配置")
+        except ImportError:
+            logger.warning("无法导入配置模块，使用默认配置")
+
+
+            # 使用默认配置的实现
+            def get_model_config(platform: str, model_type: str):
+                return {
+                    "api_key": "AIzaSyAVE8yTaPtN-TxCCHTc9Jb-aCV-Xo1EFuU",
+                    "model": "gemini-2.0-flash"
+                }
+except Exception as e:
+    logger.error(f"导入配置时出错: {str(e)}，使用默认配置")
+
+
+    # 使用默认配置的实现
+    def get_model_config(platform: str, model_type: str):
+        return {
+            "api_key": "AIzaSyAVE8yTaPtN-TxCCHTc9Jb-aCV-Xo1EFuU",
+            "model": "gemini-2.0-flash"
+        }
+
+class ChatBot:
+    def __init__(self, platform: str = "Gemini", model_type: str = "gemini-2.0-flash"):
+        """初始化聊天机器人
+
+        Args:
+            platform: 平台名称，默认为Gemini
+            model_type: 要使用的模型类型，默认为gemini-2.0-flash
+        """
+        try:
+            # 从配置获取API配置
+            config = get_model_config(platform, model_type)
+            self.api_key = config["api_key"]
+            self.model = config["model"]  # 直接使用配置中的模型名称
+
+            logger.info(f"初始化ChatBot，使用平台: {platform}, 模型: {self.model}")
+
+            google_search_tool = Tool(
+                google_search=GoogleSearch()
+            )
+
+            self.client = genai.Client(api_key=self.api_key)
+            
+            # 系统提示语
+            self.system_instruction = """你是一位经验丰富的专业投资经理，擅长基本面分析和投资决策。你的分析特点如下：
+
+                            1. 分析风格：
+                            - 专业、客观、理性
+                            - 注重数据支撑
+                            - 关注风险控制
+                            - 重视投资性价比
+
+                            2. 分析框架：
+                            - 公司基本面分析
+                            - 行业竞争格局
+                            - 估值水平评估
+                            - 风险因素识别
+                            - 投资机会判断
+
+                            3. 输出要求：
+                            - 简明扼要，重点突出
+                            - 逻辑清晰，层次分明
+                            - 数据准确，论据充分
+                            - 结论明确，建议具体
+
+                            请用专业投资经理的视角，对股票进行深入分析和投资建议。如果信息不足，请明确指出。"""
+            
+            # 对话历史
+            self.conversation_history = []
+        except Exception as e:
+            logger.error(f"初始化ChatBot时出错: {str(e)}")
+            raise
+
+    def chat(self, user_input: str, analysis_type: Optional[str] = None, temperature: float = 0.7, top_p: float = 0.7, max_tokens: int = 4096) -> dict:
+        """处理用户输入并返回AI回复
+        
+        Args:
+            user_input: 用户输入的问题
+            analysis_type: 分析类型，如company_profile或business_scale等
+            temperature: 控制输出的随机性，范围0-2，默认0.7
+            top_p: 控制输出的多样性，范围0-1，默认0.7
+            max_tokens: 控制输出的最大长度，默认4096
+            
+        Returns:
+            dict: 包含AI回复的字典
+        """
+        try:
+            from google.genai import types
+            google_search_tool = Tool(
+                google_search=GoogleSearch()
+            )
+            
+            # 确定分析类型和模型
+            model_class = None
+            if analysis_type and analysis_type in ANALYSIS_TYPE_MODEL_MAP:
+                model_class = ANALYSIS_TYPE_MODEL_MAP[analysis_type]
+                parser = JsonOutputParser(pydantic_object=model_class)
+                format_instructions = parser.get_format_instructions()
+                
+                # 在原有基础上添加格式指令
+                combined_instruction = f"{self.system_instruction}\n\n{format_instructions}"
+            else:
+                combined_instruction = self.system_instruction
+                parser = None
+            
+            response = self.client.models.generate_content_stream(
+                model=self.model,
+                config=GenerateContentConfig(
+                    system_instruction=combined_instruction,
+                    tools=[google_search_tool],
+                    response_modalities=["TEXT"],
+                    temperature=temperature,
+                    top_p=top_p,
+                    max_output_tokens=max_tokens
+                ),
+                contents=user_input
+            )
+            
+            full_response = ""
+            for chunk in response:
+                full_response += chunk.text
+                print(chunk.text, end="")
+            
+            # 如果指定了分析类型，尝试解析输出
+            if parser:
+                try:
+                    # 检查分析类型是文本分析还是数值分析
+                    is_text_analysis = model_class == TextAnalysisResult
+                    is_numerical_analysis = model_class == NumericalAnalysisResult
+                    
+                    # 使用解析方法尝试处理JSON响应
+                    json_result = self.parse_json_response(full_response, model_class)
+                    if json_result:
+                        return json_result
+                    
+                    # 如果使用解析方法失败，尝试使用parser解析
+                    parsed_result = parser.parse(full_response)
+                    structured_result = parsed_result.dict()
+                    
+                    if isinstance(parsed_result, TextAnalysisResult):
+                        # 文本分析结果
+                        return {
+                            "response": structured_result["analysis_text"],
+                            "reasoning_process": structured_result.get("reasoning_process"),
+                            "references": structured_result.get("references")
+                        }
+                    elif isinstance(parsed_result, NumericalAnalysisResult):
+                        # 数值分析结果
+                        return {
+                            "response": structured_result["value"],
+                            "description": structured_result.get("description", "")
+                        }
+                except Exception as e:
+                    logger.warning(f"解析响应失败: {str(e)}，返回原始响应")
+                    
+                    # 最后尝试根据分析类型直接构建结果
+                    if is_text_analysis:
+                        return {
+                            "response": full_response,
+                            "reasoning_process": None,
+                            "references": None
+                        }
+                    elif is_numerical_analysis:
+                        try:
+                            # 尝试从文本中提取数字
+                            import re
+                            number_match = re.search(r'\b(\d+)\b', full_response)
+                            value = int(number_match.group(1)) if number_match else 0
+                            return {
+                                "response": value,
+                                "description": full_response
+                            }
+                        except:
+                            return {
+                                "response": 0,
+                                "description": full_response
+                            }
+                    else:
+                        return {
+                            "response": full_response,
+                            "reasoning_process": None,
+                            "references": None
+                        }
+            else:
+                # 普通响应
+                return {
+                    "response": full_response,
+                    "reasoning_process": None,
+                    "references": None
+                }
+
+        except Exception as e:
+            logger.error(f"聊天失败: {str(e)}")
+            error_msg = f"抱歉，发生错误: {str(e)}"
+            print(f"\n{error_msg}")
+            return {"response": error_msg, "reasoning_process": None, "references": None}
+
+    def clear_history(self):
+        """清除对话历史"""
+        self.conversation_history = []
+        # 重置会话
+        if hasattr(self, 'model_instance'):
+            self.chat_session = self.model_instance.start_chat()
+        print("对话历史已清除")
+
+    def run(self):
+        """运行聊天机器人"""
+        print("欢迎使用Gemini AI助手！输入 'quit' 退出，输入 'clear' 清除对话历史。")
+        print("-" * 50)
+
+        while True:
+            try:
+                # 获取用户输入
+                user_input = input("\n你: ").strip()
+
+                # 检查退出命令
+                if user_input.lower() == 'quit':
+                    print("感谢使用，再见！")
+                    break
+
+                # 检查清除历史命令
+                if user_input.lower() == 'clear':
+                    self.clear_history()
+                    continue
+
+                # 如果输入为空，继续下一轮
+                if not user_input:
+                    continue
+
+                # 获取AI回复
+                self.chat(user_input)
+                print("-" * 50)
+
+            except KeyboardInterrupt:
+                print("\n感谢使用，再见！")
+                break
+            except Exception as e:
+                logger.error(f"运行错误: {str(e)}")
+                print(f"发生错误: {str(e)}")
+
+    def parse_json_response(self, full_response: str, model_class) -> dict:
+        """解析JSON响应，处理不同的JSON结构
+        
+        Args:
+            full_response: 完整的响应文本
+            model_class: 模型类（TextAnalysisResult或NumericalAnalysisResult）
+            
+        Returns:
+            dict: 包含解析结果的字典
+        """
+        # 检查分析类型
+        is_text_analysis = model_class == TextAnalysisResult
+        is_numerical_analysis = model_class == NumericalAnalysisResult
+        
+        try:
+            # 尝试提取JSON内容
+            json_text = full_response
+            # 如果文本以```json开头且以```结尾，则提取中间部分
+            if "```json" in json_text and "```" in json_text.split("```json", 1)[1]:
+                json_text = json_text.split("```json", 1)[1].split("```", 1)[0].strip()
+            # 如果文本以```开头且以```结尾，则提取中间部分
+            elif "```" in json_text and "```" in json_text.split("```", 1)[1]:
+                json_text = json_text.split("```", 1)[1].split("```", 1)[0].strip()
+            
+            # 尝试解析JSON
+            json_obj = json.loads(json_text)
+            
+            # 情况1: 直接包含字段的格式 {"analysis_text": "...", "reasoning_process": null, "references": [...]}
+            if is_text_analysis and "analysis_text" in json_obj:
+                return {
+                    "response": json_obj["analysis_text"],
+                    "reasoning_process": json_obj.get("reasoning_process"),
+                    "references": json_obj.get("references", [])
+                }
+            elif is_numerical_analysis and "value" in json_obj:
+                value = json_obj["value"]
+                if isinstance(value, str) and value.isdigit():
+                    value = int(value)
+                elif isinstance(value, (int, float)):
+                    value = int(value)
+                else:
+                    value = 0
+                    
+                return {
+                    "response": value,
+                    "description": json_obj.get("description", "")
+                }
+            
+            # 情况2: properties中包含字段的格式
+            if "properties" in json_obj and isinstance(json_obj["properties"], dict):
+                properties = json_obj["properties"]
+                
+                if is_text_analysis and "analysis_text" in properties:
+                    return {
+                        "response": properties["analysis_text"],
+                        "reasoning_process": properties.get("reasoning_process"),
+                        "references": properties.get("references", [])
+                    }
+                elif is_numerical_analysis and "value" in properties:
+                    value = properties["value"]
+                    if isinstance(value, str) and value.isdigit():
+                        value = int(value)
+                    elif isinstance(value, (int, float)):
+                        value = int(value)
+                    else:
+                        value = 0
+                        
+                    return {
+                        "response": value,
+                        "description": properties.get("description", "")
+                    }
+            
+            # 如果无法识别结构，使用parser解析
+            raise ValueError("无法识别的JSON结构，尝试使用parser解析")
+            
+        except Exception as e:
+            # 记录错误但不抛出，让调用方法继续尝试其他解析方式
+            logger.warning(f"JSON解析失败: {str(e)}")
+            return None
+
+if __name__ == "__main__":
+    # 设置日志级别
+    logging.basicConfig(
+        level=logging.INFO,
+        format='%(asctime)s - %(name)s - %(levelname)s - %(message)s'
+    )
+
+    # 创建并运行聊天机器人
+    bot = ChatBot()
+    bot.run() 

src/fundamentals_llm/chat_bot_with_offline.py

@@ -3,7 +3,7 @@ from openai import OpenAI
 import os
 import sys
 import time
-import random
+
 
 # 设置日志记录
 logger = logging.getLogger(__name__)
@@ -102,8 +102,19 @@ class ChatBot:
             logger.error(f"初始化ChatBot时出错: {str(e)}")
             raise
 
-    def chat(self, user_input: str) -> str:
-        """处理用户输入并返回AI回复"""
+    def chat(self, user_input: str, temperature: float = 1.0, top_p: float = 0.7, max_tokens: int = 4096, frequency_penalty: float = 0.0) -> str:
+        """处理用户输入并返回AI回复
+        
+        Args:
+            user_input: 用户输入的问题
+            temperature: 控制输出的随机性，范围0-2，默认1.0
+            top_p: 控制输出的多样性，范围0-1，默认0.7
+            max_tokens: 控制输出的最大长度，默认4096
+            frequency_penalty: 控制重复惩罚，范围-2到2，默认0.0
+            
+        Returns:
+            str: AI的回答内容
+        """
         try:
             # # 添加用户消息到对话历史-多轮
             self.conversation_history.append({
@@ -116,7 +127,10 @@ class ChatBot:
             stream = self.client.chat.completions.create(
                 model=self.model,
                 messages=self.conversation_history,
-                temperature=0,
+                temperature=temperature,
+                top_p=top_p,
+                max_tokens=max_tokens,
+                frequency_penalty=frequency_penalty,
                 stream=True,
                 timeout=600
             )

src/fundamentals_llm/enterprise_screener.py

@@ -130,9 +130,15 @@ class EnterpriseScreener:
             {
                 'dimension': 'investment_advice',
                 'field': 'investment_advice_type',
-                'operator': '!=',
-                'value': '不建议'
-            }
+                'operator': 'in',
+                'value': ["'短期'","'中期'","'长期'"]
+            },
+            {
+                'dimension': 'financial_report',
+                'field': 'financial_report_level',
+                'operator': '>=',
+                'value': -1
+            },
         ]
         return self._screen_stocks_by_conditions(conditions, limit)
 

src/fundamentals_llm/fundamental_analysis.py

@@ -1,7 +1,11 @@
 import logging
 import os
-from datetime import datetime
-from typing import Dict, List, Optional, Tuple, Callable
+import sys
+from datetime import datetime, timedelta
+import time
+import redis
+from typing import Dict, List, Optional, Tuple, Callable, Any
+
 # 修改导入路径，使用相对导入
 try:
     # 尝试相对导入
@@ -9,6 +13,7 @@ try:
     from .chat_bot_with_offline import ChatBot as OfflineChatBot
     from .fundamental_analysis_database import get_db, save_analysis_result, update_analysis_result, get_analysis_result
     from .pdf_generator import PDFGenerator
+    from .text_processor import TextProcessor
 except ImportError:
     # 如果相对导入失败，尝试尝试绝对导入
     try:
@@ -16,18 +21,17 @@ except ImportError:
         from src.fundamentals_llm.chat_bot_with_offline import ChatBot as OfflineChatBot
         from src.fundamentals_llm.fundamental_analysis_database import get_db, save_analysis_result, update_analysis_result, get_analysis_result
         from src.fundamentals_llm.pdf_generator import PDFGenerator
+        from src.fundamentals_llm.text_processor import TextProcessor
     except ImportError:
         # 最后尝试直接导入（适用于当前目录已在PYTHONPATH中的情况）
         from chat_bot import ChatBot
         from chat_bot_with_offline import ChatBot as OfflineChatBot
         from fundamental_analysis_database import get_db, save_analysis_result, update_analysis_result, get_analysis_result
         from pdf_generator import PDFGenerator
+        from text_processor import TextProcessor
 import json
 import re
 
-# 设置日志记录
-logger = logging.getLogger(__name__)
-
 # 获取项目根目录的绝对路径
 ROOT_DIR = os.path.dirname(os.path.dirname(os.path.abspath(__file__)))
 
@@ -58,6 +62,34 @@ logging.basicConfig(
     datefmt=date_format
 )
 
+logger = logging.getLogger(__name__)
+logger.info("测试日志输出 - 程序启动")
+
+from typing import Dict, List, Optional, Any, Union
+from pydantic import BaseModel, Field
+
+# 定义基础数据结构
+class TextAnalysisResult(BaseModel):
+    """文本分析结果，包含分析正文、推理过程和引用URL"""
+    analysis_text: str = Field(description="详细的分析文本")
+    reasoning_process: Optional[str] = Field(description="模型的推理过程", default=None)
+    references: Optional[List[str]] = Field(description="参考资料和引用URL列表", default=None)
+
+class NumericalAnalysisResult(BaseModel):
+    """数值分析结果，包含数值和分析描述"""
+    value: str = Field(description="评估值")
+    description: str = Field(description="评估描述")
+
+# 添加Redis客户端
+redis_client = redis.Redis(
+    host='192.168.18.208',  # Redis服务器地址，根据实际情况调整
+    port=6379,
+    password='wlkj2018',
+    db=14,
+    socket_timeout=5,
+    decode_responses=True
+)
+
 class FundamentalAnalyzer:
     """基本面分析器"""
     
@@ -66,9 +98,10 @@ class FundamentalAnalyzer:
         # 使用联网模型进行基本面分析
         self.chat_bot = ChatBot(model_type="online_bot")
         # 使用离线模型进行其他分析
-        self.offline_bot = OfflineChatBot(platform="tl_private", model_type="ds-v1")
+        self.offline_bot = OfflineChatBot(platform="volc", model_type="offline_model")
         # 千问打杂
-        self.offline_bot_tl_qw = OfflineChatBot(platform="tl_qw_private", model_type="qwq")
+        # self.offline_bot_tl_qw = OfflineChatBot(platform="tl_qw_private", model_type="qwq")
+        self.offline_bot_tl_qw = OfflineChatBot(platform="tl_qw_private", model_type="GLM")
         self.db = next(get_db())
         
         # 定义维度映射
@@ -161,6 +194,8 @@ class FundamentalAnalyzer:
             - 关键战略决策和转型
 
             请提供专业、客观的分析，突出关键信息，避免冗长描述。"""
+            #开头清上下文缓存
+            self.chat_bot.clear_history()
             # 获取AI分析结果
             result = self.chat_bot.chat(prompt)
             
@@ -554,8 +589,8 @@ class FundamentalAnalyzer:
                     请仅返回一个数值：2、1、0或-1，不要包含任何解释或说明。"""
             self.offline_bot_tl_qw.clear_history()
             # 使用离线模型进行分析
-            space_value_str = self.offline_bot_tl_qw.chat(prompt)
-            space_value_str = self._clean_model_output(space_value_str)
+            space_value_str = self.offline_bot_tl_qw.chat(prompt,temperature=0.0)
+            space_value_str = TextProcessor.clean_thought_process(space_value_str)
             # 提取数值
             space_value = 0  # 默认值
             
@@ -658,9 +693,9 @@ class FundamentalAnalyzer:
                     请仅返回一个数值：1、0或-1，不要包含任何解释或说明。"""
             self.offline_bot_tl_qw.clear_history()
             # 使用离线模型进行分析
-            events_value_str = self.offline_bot_tl_qw.chat(prompt)
+            events_value_str = self.offline_bot_tl_qw.chat(prompt,temperature=0.0)
             # 数据清洗
-            events_value_str = self._clean_model_output(events_value_str)
+            events_value_str = TextProcessor.clean_thought_process(events_value_str)
             # 提取数值
             events_value = 0  # 默认值
             
@@ -700,14 +735,14 @@ class FundamentalAnalyzer:
     def analyze_stock_discussion(self, stock_code: str, stock_name: str) -> bool:
         """分析股吧讨论内容"""
         try:
-            prompt = f"""请对{stock_name}({stock_code})的股吧讨论内容进行简要分析，要求输出控制在300字以内，请严格按照以下格式输出：
+            prompt = f"""请对{stock_name}({stock_code})的股吧讨论内容进行简要分析，要求输出控制在400字以内(主要讨论话题200字，重要信息汇总200字)，请严格按照以下格式输出：
 
-                    1. 主要讨论话题（150字左右）：
+                    1. 主要讨论话题：
                     - 近期热点事件
                     - 投资者关注焦点
                     - 市场情绪倾向
                     
-                    2. 重要信息汇总（150字左右）：
+                    2. 重要信息汇总：
                     - 公司相关动态
                     - 行业政策变化
                     - 市场预期变化
@@ -762,10 +797,10 @@ class FundamentalAnalyzer:
                     请仅返回一个数值：1、0或-1，不要包含任何解释或说明。"""
             self.offline_bot_tl_qw.clear_history()
             # 使用离线模型进行分析
-            emotion_value_str = self.offline_bot_tl_qw.chat(prompt)
+            emotion_value_str = self.offline_bot_tl_qw.chat(prompt,temperature=0.0)
 
             # 数据清洗
-            emotion_value_str = self._clean_model_output(emotion_value_str)
+            emotion_value_str = TextProcessor.clean_thought_process(emotion_value_str)
             # 提取数值
             emotion_value = 0  # 默认值
             
@@ -1015,10 +1050,10 @@ class FundamentalAnalyzer:
                     
                     只需要输出一个数值，不要输出任何说明或解释。只输出：2,1,0,-1或-2。"""
             
-            response = self.chat_bot.chat(prompt)
+            response = self.offline_bot_tl_qw.chat(prompt,temperature=0.0)
             
             # 提取数值
-            rating_str = self._extract_numeric_value_from_response(response)
+            rating_str = TextProcessor.extract_numeric_value_from_response(response)
             
             # 尝试将响应转换为整数
             try:
@@ -1057,10 +1092,10 @@ class FundamentalAnalyzer:
                         
                         只需要输出一个数值，不要输出任何说明或解释。只输出：1、0、-1，不要包含任何解释或说明。"""
             
-            response = self.chat_bot.chat(prompt)
+            response = self.offline_bot_tl_qw.chat(prompt,temperature=0.0)
             
             # 提取数值
-            odds_str = self._extract_numeric_value_from_response(response)
+            odds_str = TextProcessor.extract_numeric_value_from_response(response)
             
             # 尝试将响应转换为整数
             try:
@@ -1200,8 +1235,8 @@ class FundamentalAnalyzer:
                     只需要输出一个数值，不要输出任何说明或解释。只输出：-1、0或1。"""
             
             self.offline_bot_tl_qw.clear_history()
-            response = self.offline_bot_tl_qw.chat(prompt)
-            pe_hist_str = self._clean_model_output(response)
+            response = self.offline_bot_tl_qw.chat(prompt,temperature=0.0)
+            pe_hist_str =  TextProcessor.clean_thought_process(response)
             
             try:
                 pe_hist = int(pe_hist_str)
@@ -1240,8 +1275,8 @@ class FundamentalAnalyzer:
                     只需要输出一个数值，不要输出任何说明或解释。只输出：-1、0或1。"""
             
             self.offline_bot_tl_qw.clear_history()
-            response = self.offline_bot_tl_qw.chat(prompt)
-            pb_hist_str = self._clean_model_output(response)
+            response = self.offline_bot_tl_qw.chat(prompt,temperature=0.0)
+            pb_hist_str = TextProcessor.clean_thought_process(response)
             
             try:
                 pb_hist = int(pb_hist_str)
@@ -1280,8 +1315,8 @@ class FundamentalAnalyzer:
                     只需要输出一个数值，不要输出任何说明或解释。只输出：-1、0或1。"""
             
             self.offline_bot_tl_qw.clear_history()
-            response = self.offline_bot_tl_qw.chat(prompt)
-            pe_ind_str = self._clean_model_output(response)
+            response = self.offline_bot_tl_qw.chat(prompt,temperature=0.0)
+            pe_ind_str = TextProcessor.clean_thought_process(response)
             
             try:
                 pe_ind = int(pe_ind_str)
@@ -1320,8 +1355,8 @@ class FundamentalAnalyzer:
                     只需要输出一个数值，不要输出任何说明或解释。只输出：-1、0或1。"""
             
             self.offline_bot_tl_qw.clear_history()
-            response = self.offline_bot_tl_qw.chat(prompt)
-            pb_ind_str = self._clean_model_output(response)
+            response = self.offline_bot_tl_qw.chat(prompt,temperature=0.0)
+            pb_ind_str = TextProcessor.clean_thought_process(response)
             
             try:
                 pb_ind = int(pb_ind_str)
@@ -1373,9 +1408,9 @@ class FundamentalAnalyzer:
             {json.dumps(all_results, ensure_ascii=False, indent=2)}"""
             self.offline_bot.clear_history()
             # 使用离线模型生成建议
-            result = self.offline_bot.chat(prompt)
+            result = self.offline_bot.chat(prompt,max_tokens=20000)
             # 清理模型输出
-            result = self._clean_model_output(result)
+            result = TextProcessor.clean_thought_process(result)
             # 保存到数据库
             success = save_analysis_result(
                 self.db,
@@ -1441,93 +1476,6 @@ class FundamentalAnalyzer:
             logger.error(f"提取投资建议类型失败: {str(e)}")
             return None
         
-    def _clean_model_output(self, output: str) -> str:
-        """清理模型输出，移除推理过程，只保留最终结果
-        
-        Args:
-            output: 模型原始输出文本
-            
-        Returns:
-            str: 清理后的输出文本
-        """
-        try:
-            # 找到</think>标签的位置
-            think_end = output.find('</think>')
-            if think_end != -1:
-                # 移除</think>标签及其之前的所有内容
-                output = output[think_end + len('</think>'):]
-            
-            # 处理可能存在的空行
-            lines = output.split('\n')
-            cleaned_lines = []
-            for line in lines:
-                line = line.strip()
-                if line:  # 只保留非空行
-                    cleaned_lines.append(line)
-            
-            # 重新组合文本
-            output = '\n'.join(cleaned_lines)
-            
-            return output.strip()
-            
-        except Exception as e:
-            logger.error(f"清理模型输出失败: {str(e)}")
-            return output.strip()
-
-    def _extract_numeric_value_from_response(self, response: str) -> str:
-        """从模型响应中提取数值，移除参考资料和推理过程
-        
-        Args:
-            response: 模型原始响应文本或响应对象
-            
-        Returns:
-            str: 提取的数值字符串
-        """
-        try:
-            # 处理响应对象（包含response字段的字典）
-            if isinstance(response, dict) and "response" in response:
-                response = response["response"]
-                
-            # 确保响应是字符串
-            if not isinstance(response, str):
-                logger.warning(f"响应不是字符串类型: {type(response)}")
-                return "0"
-                
-            # 移除推理过程部分
-            reasoning_start = response.find("推理过程：")
-            if reasoning_start != -1:
-                response = response[:reasoning_start].strip()
-                
-            # 移除参考资料部分（通常以 [数字] 开头的行）
-            lines = response.split("\n")
-            cleaned_lines = []
-            
-            for line in lines:
-                # 跳过参考资料行（通常以 [数字] 开头）
-                if re.match(r'\[\d+\]', line.strip()):
-                    continue
-                cleaned_lines.append(line)
-                
-            response = "\n".join(cleaned_lines).strip()
-            
-            # 提取数值
-            # 先尝试直接将整个响应转换为数值
-            if response.strip() in ["-2", "-1", "0", "1", "2"]:
-                return response.strip()
-                
-            # 如果整个响应不是数值，尝试匹配第一个数值
-            match = re.search(r'([-]?[0-9])', response)
-            if match:
-                return match.group(1)
-                
-            # 如果没有找到数值，返回默认值
-            logger.warning(f"未能从响应中提取数值: {response}")
-            return "0"
-            
-        except Exception as e:
-            logger.error(f"从响应中提取数值失败: {str(e)}")
-            return "0"
-
     def _try_extract_advice_type(self, advice_text: str, max_attempts: int = 3) -> Optional[str]:
         """尝试多次从投资建议中提取建议类型
         
@@ -1549,7 +1497,7 @@ class FundamentalAnalyzer:
                 
                 # 使用千问离线模型提取建议类型
 
-                result = self.offline_bot_tl_qw.chat(prompt)
+                result = self.offline_bot_tl_qw.chat(prompt,temperature=0.0)
                 
                 # 检查是否是错误响应
                 if isinstance(result, str) and "抱歉，发生错误" in result:
@@ -1557,7 +1505,7 @@ class FundamentalAnalyzer:
                     continue
                 
                 # 清理模型输出
-                cleaned_result = self._clean_model_output(result)
+                cleaned_result = TextProcessor.clean_thought_process(result)
                 
                 # 检查结果是否为有效类型
                 if cleaned_result in valid_types:
@@ -1644,6 +1592,24 @@ class FundamentalAnalyzer:
             Optional[str]: 生成的PDF文件路径，如果失败则返回None
         """
         try:
+            # 检查是否已存在PDF文件
+            reports_dir = os.path.join(os.path.dirname(os.path.dirname(__file__)), 'reports')
+            os.makedirs(reports_dir, exist_ok=True)
+            
+            # 构建可能的文件名格式
+            possible_filenames = [
+                f"{stock_name}_{stock_code}_analysis.pdf",
+                f"{stock_name}_{stock_code}.SZ_analysis.pdf",
+                f"{stock_name}_{stock_code}.SH_analysis.pdf"
+            ]
+            
+            # 检查是否存在已生成的PDF文件
+            for filename in possible_filenames:
+                filepath = os.path.join(reports_dir, filename)
+                if os.path.exists(filepath):
+                    logger.info(f"找到已存在的PDF报告: {filepath}")
+                    return filepath
+            
             # 维度名称映射
             dimension_names = {
                 "company_profile": "公司简介",
@@ -1669,27 +1635,156 @@ class FundamentalAnalyzer:
                 logger.warning(f"未找到 {stock_name}({stock_code}) 的任何分析结果")
                 return None
             
+            # 确保字体目录存在
+            fonts_dir = os.path.join(os.path.dirname(__file__), "fonts")
+            os.makedirs(fonts_dir, exist_ok=True)
+            
+            # 检查是否存在字体文件，如果不存在则创建简单的默认字体标记文件
+            font_path = os.path.join(fonts_dir, "simhei.ttf")
+            if not os.path.exists(font_path):
+                # 尝试从系统字体目录复制
+                try:
+                    import shutil
+                    # 尝试常见的系统字体位置
+                    system_fonts = [
+                        "C:/Windows/Fonts/simhei.ttf",  # Windows
+                        "/usr/share/fonts/truetype/wqy/wqy-microhei.ttc",  # Linux
+                        "/usr/share/fonts/wqy-microhei/wqy-microhei.ttc",  # 其他Linux
+                        "/System/Library/Fonts/PingFang.ttc"  # macOS
+                    ]
+                    
+                    for system_font in system_fonts:
+                        if os.path.exists(system_font):
+                            shutil.copy2(system_font, font_path)
+                            logger.info(f"已复制字体文件: {system_font} -> {font_path}")
+                            break
+                except Exception as font_error:
+                    logger.warning(f"复制字体文件失败: {str(font_error)}")
+            
             # 创建PDF生成器实例
             generator = PDFGenerator()
             
             # 生成PDF报告
+            try:
+                # 第一次尝试生成
                 filepath = generator.generate_pdf(
                     title=f"{stock_name}({stock_code}) 基本面分析报告",
                     content_dict=content_dict,
+                    output_dir=reports_dir,
                     filename=f"{stock_name}_{stock_code}_analysis.pdf"
                 )
                 
                 if filepath:
                     logger.info(f"PDF报告已生成: {filepath}")
-            else:
-                logger.error("PDF报告生成失败")
-            
                     return filepath
+            except Exception as pdf_error:
+                logger.error(f"生成PDF报告第一次尝试失败: {str(pdf_error)}")
+                
+                # 如果是字体问题，可能需要使用备选方案
+                if "找不到中文字体文件" in str(pdf_error):
+                    # 导出为文本文件作为备选
+                    try:
+                        txt_filename = f"{stock_name}_{stock_code}_analysis.txt"
+                        txt_filepath = os.path.join(reports_dir, txt_filename)
+                        
+                        with open(txt_filepath, 'w', encoding='utf-8') as f:
+                            f.write(f"{stock_name}({stock_code}) 基本面分析报告\n")
+                            f.write(f"生成时间：{datetime.now().strftime('%Y年%m月%d日 %H:%M:%S')}\n\n")
+                            
+                            for section_title, content in content_dict.items():
+                                if content:
+                                    f.write(f"## {section_title}\n\n")
+                                    f.write(f"{content}\n\n")
+                        
+                        logger.info(f"由于PDF生成失败，已生成文本报告: {txt_filepath}")
+                        return txt_filepath
+                    except Exception as txt_error:
+                        logger.error(f"生成文本报告失败: {str(txt_error)}")
+            
+            logger.error("PDF报告生成失败")
+            return None
             
         except Exception as e:
             logger.error(f"生成PDF报告失败: {str(e)}")
             return None
 
+    def is_stock_locked(self, stock_code: str, dimension: str) -> bool:
+        """检查股票是否已被锁定
+        
+        Args:
+            stock_code: 股票代码
+            dimension: 分析维度
+            
+        Returns:
+            bool: 是否被锁定
+        """
+        try:
+            lock_key = f"stock_analysis_lock:{stock_code}:{dimension}"
+            
+            # 检查是否存在锁
+            existing_lock = redis_client.get(lock_key)
+            if existing_lock:
+                lock_time = int(existing_lock)
+                current_time = int(time.time())
+                
+                # 锁超过30分钟(1800秒)视为过期
+                if current_time - lock_time > 1800:
+                    # 锁已过期，可以释放
+                    redis_client.delete(lock_key)
+                    logger.info(f"股票 {stock_code} 维度 {dimension} 的过期锁已释放")
+                    return False
+                else:
+                    # 锁未过期，被锁定
+                    return True
+            
+            # 不存在锁
+            return False
+        except Exception as e:
+            logger.error(f"检查股票 {stock_code} 锁状态时出错: {str(e)}")
+            # 出错时保守处理，返回未锁定
+            return False
+    
+    def lock_stock(self, stock_code: str, dimension: str) -> bool:
+        """锁定股票
+        
+        Args:
+            stock_code: 股票代码
+            dimension: 分析维度
+            
+        Returns:
+            bool: 是否成功锁定
+        """
+        try:
+            lock_key = f"stock_analysis_lock:{stock_code}:{dimension}"
+            current_time = int(time.time())
+            
+            # 设置锁，过期时间1小时
+            redis_client.set(lock_key, current_time, ex=3600)
+            logger.info(f"股票 {stock_code} 维度 {dimension} 已锁定")
+            return True
+        except Exception as e:
+            logger.error(f"锁定股票 {stock_code} 时出错: {str(e)}")
+            return False
+    
+    def unlock_stock(self, stock_code: str, dimension: str) -> bool:
+        """解锁股票
+        
+        Args:
+            stock_code: 股票代码
+            dimension: 分析维度
+            
+        Returns:
+            bool: 是否成功解锁
+        """
+        try:
+            lock_key = f"stock_analysis_lock:{stock_code}:{dimension}"
+            redis_client.delete(lock_key)
+            logger.info(f"股票 {stock_code} 维度 {dimension} 已解锁")
+            return True
+        except Exception as e:
+            logger.error(f"解锁股票 {stock_code} 时出错: {str(e)}")
+            return False
+
 def test_single_method(method: Callable, stock_code: str, stock_name: str) -> bool:
     """测试单个分析方法"""
     try:
@@ -1713,11 +1808,6 @@ def test_single_stock(analyzer: FundamentalAnalyzer, stock_code: str, stock_name
 
 def main():
     """主函数"""
-    # 设置日志级别
-    logging.basicConfig(
-        level=logging.INFO,
-        format='%(asctime)s - %(name)s - %(levelname)s - %(message)s'
-    )
 
     # 测试股票列表
     test_stocks = [

src/fundamentals_llm/pdf_generator.py

@@ -11,25 +11,25 @@ import markdown2
 from bs4 import BeautifulSoup
 import os
 from datetime import datetime
-from fpdf import FPDF
-import matplotlib.pyplot as plt
+import shutil
+
 import matplotlib
 matplotlib.use('Agg')
 
 # 修改导入路径，使用相对导入
 try:
     # 尝试相对导入
-    from .chat_bot_with_offline import ChatBot
-    from .fundamental_analysis_database import get_db, get_analysis_result
+    from .chat_bot import ChatBot
+    from .chat_bot_with_offline import ChatBot as OfflineChatBot
 except ImportError:
     # 如果相对导入失败，尝试绝对导入
     try:
-        from src.fundamentals_llm.chat_bot_with_offline import ChatBot
-        from src.fundamentals_llm.fundamental_analysis_database import get_db, get_analysis_result
+        from src.fundamentals_llm.chat_bot import ChatBot
+        from src.fundamentals_llm.chat_bot_with_offline import ChatBot as OfflineChatBot
     except ImportError:
         # 最后尝试直接导入
-        from chat_bot_with_offline import ChatBot
-        from fundamental_analysis_database import get_db, get_analysis_result
+        from chat_bot import ChatBot
+        from chat_bot_with_offline import ChatBot as OfflineChatBot
 
 # 设置日志记录
 logger = logging.getLogger(__name__)
@@ -39,30 +39,9 @@ class PDFGenerator:
     
     def __init__(self):
         """初始化PDF生成器"""
-        # 注册中文字体
-        try:
-            # 尝试使用系统自带的中文字体
-            if os.name == 'nt':  # Windows
-                font_path = "C:/Windows/Fonts/simhei.ttf"  # 黑体
-            else:  # Linux/Mac
-                font_path = "/usr/share/fonts/truetype/droid/DroidSansFallback.ttf"
-            
-            if os.path.exists(font_path):
-                pdfmetrics.registerFont(TTFont('SimHei', font_path))
-                self.font_name = 'SimHei'
-            else:
-                # 如果找不到系统字体，尝试使用当前目录下的字体
-                font_path = os.path.join(os.path.dirname(__file__), "fonts", "simhei.ttf")
-                if os.path.exists(font_path):
-                    pdfmetrics.registerFont(TTFont('SimHei', font_path))
-                    self.font_name = 'SimHei'
-                else:
-                    raise FileNotFoundError("找不到中文字体文件")
-        except Exception as e:
-            logger.error(f"注册中文字体失败: {str(e)}")
-            raise
-        
-        self.chat_bot = ChatBot()
+        # 尝试注册中文字体
+        self.font_name = self._register_chinese_font()
+        self.chat_bot = OfflineChatBot(platform="volc", model_type="offline_model")  # 使用离线模式
         self.styles = getSampleStyleSheet()
         
         # 创建自定义样式
@@ -139,6 +118,64 @@ class PDFGenerator:
             textColor=colors.HexColor('#333333')
         ))
     
+    def _register_chinese_font(self):
+        """查找并注册中文字体
+        
+        Returns:
+            str: 注册的字体名称，如果失败则使用默认字体
+        """
+        try:
+            # 可能的字体文件位置列表
+            font_locations = [
+                # 当前目录/子目录字体位置
+                os.path.join(os.path.dirname(__file__), "fonts", "simhei.ttf"),
+                os.path.join(os.path.dirname(__file__), "fonts", "wqy-microhei.ttc"),
+                os.path.join(os.path.dirname(__file__), "fonts", "wqy-zenhei.ttc"),
+                # Docker中可能的位置
+                "/app/src/fundamentals_llm/fonts/simhei.ttf",
+                # Windows系统字体位置
+                "C:/Windows/Fonts/simhei.ttf",
+                "C:/Windows/Fonts/simfang.ttf",
+                "C:/Windows/Fonts/simsun.ttc",
+                # Linux/Mac系统字体位置
+                "/usr/share/fonts/truetype/wqy/wqy-microhei.ttc",
+                "/usr/share/fonts/truetype/wqy/wqy-zenhei.ttc",
+                "/usr/share/fonts/wqy-microhei/wqy-microhei.ttc",
+                "/usr/share/fonts/wqy-zenhei/wqy-zenhei.ttc",
+                "/usr/share/fonts/truetype/droid/DroidSansFallback.ttf",
+                "/usr/share/fonts/opentype/noto/NotoSansCJK-Regular.ttc",
+                "/System/Library/Fonts/PingFang.ttc"  # macOS
+            ]
+            
+            # 尝试各个位置
+            for font_path in font_locations:
+                if os.path.exists(font_path):
+                    logger.info(f"使用字体文件: {font_path}")
+                    # 为防止字体文件名称不同，统一拷贝到字体目录并重命名
+                    font_dir = os.path.join(os.path.dirname(__file__), "fonts")
+                    os.makedirs(font_dir, exist_ok=True)
+                    target_path = os.path.join(font_dir, "simhei.ttf")
+                    
+                    # 如果字体文件不在目标位置，拷贝过去
+                    if font_path != target_path and not os.path.exists(target_path):
+                        try:
+                            shutil.copy2(font_path, target_path)
+                            logger.info(f"已将字体文件复制到: {target_path}")
+                        except Exception as e:
+                            logger.warning(f"复制字体文件失败: {str(e)}")
+                    
+                    # 注册字体
+                    pdfmetrics.registerFont(TTFont('SimHei', target_path))
+                    return 'SimHei'
+            
+            # 如果所有位置都找不到，使用默认字体
+            logger.warning("找不到中文字体文件，将使用默认字体")
+            return 'Helvetica'
+            
+        except Exception as e:
+            logger.error(f"注册中文字体失败: {str(e)}")
+            return 'Helvetica'  # 使用默认字体
+    
     def _convert_markdown_to_flowables(self, markdown_text: str) -> List:
         """将Markdown文本转换为PDF流对象
         

src/fundamentals_llm/setup_fonts.py

@@ -0,0 +1,190 @@
+#!/usr/bin/env python
+# -*- coding: utf-8 -*-
+
+"""
+中文字体安装工具
+
+此脚本用于下载和安装中文字体，以支持PDF生成功能。
+"""
+
+import os
+import sys
+import logging
+import shutil
+import tempfile
+import platform
+
+# 配置日志
+logging.basicConfig(
+    level=logging.INFO,
+    format='%(asctime)s - %(name)s - %(levelname)s - %(message)s'
+)
+logger = logging.getLogger(__name__)
+
+def download_wqy_font():
+    """
+    下载文泉驿微米黑字体
+    
+    Returns:
+        str: 下载的字体文件路径，如果失败则返回None
+    """
+    try:
+        import requests
+        
+        # 创建临时目录
+        temp_dir = tempfile.mkdtemp()
+        font_file = os.path.join(temp_dir, "wqy-microhei.ttc")
+        
+        # 文泉驿微米黑字体的下载链接
+        url = "https://mirrors.tuna.tsinghua.edu.cn/osdn/wqy/wqy-microhei-0.2.0-beta.tar.gz"
+        
+        logger.info(f"开始下载文泉驿微米黑字体: {url}")
+        response = requests.get(url, stream=True)
+        response.raise_for_status()
+        
+        # 下载压缩包
+        tar_file = os.path.join(temp_dir, "wqy-microhei.tar.gz")
+        with open(tar_file, 'wb') as f:
+            for chunk in response.iter_content(chunk_size=8192):
+                f.write(chunk)
+        
+        # 解压字体文件
+        import tarfile
+        with tarfile.open(tar_file) as tar:
+            # 提取字体文件
+            for member in tar.getmembers():
+                if member.name.endswith(('.ttc', '.ttf')):
+                    member.name = os.path.basename(member.name)
+                    tar.extract(member, temp_dir)
+                    extracted_file = os.path.join(temp_dir, member.name)
+                    if os.path.exists(extracted_file):
+                        return extracted_file
+        
+        logger.error("在压缩包中未找到字体文件")
+        return None
+    
+    except Exception as e:
+        logger.error(f"下载字体失败: {str(e)}")
+        return None
+    
+def find_system_fonts():
+    """
+    在系统中查找可用的中文字体
+    
+    Returns:
+        str: 找到的字体文件路径，如果找不到则返回None
+    """
+    # 常见的中文字体位置
+    font_locations = []
+    
+    system = platform.system()
+    if system == "Windows":
+        font_locations = [
+            "C:/Windows/Fonts/simhei.ttf",
+            "C:/Windows/Fonts/simsun.ttc",
+            "C:/Windows/Fonts/msyh.ttf",
+            "C:/Windows/Fonts/simfang.ttf",
+        ]
+    elif system == "Darwin":  # macOS
+        font_locations = [
+            "/System/Library/Fonts/PingFang.ttc",
+            "/Library/Fonts/Microsoft/SimHei.ttf",
+            "/Library/Fonts/Microsoft/SimSun.ttf",
+        ]
+    else:  # Linux
+        font_locations = [
+            "/usr/share/fonts/truetype/wqy/wqy-microhei.ttc",
+            "/usr/share/fonts/wqy-microhei/wqy-microhei.ttc",
+            "/usr/share/fonts/wenquanyi/wqy-microhei/wqy-microhei.ttc",
+            "/usr/share/fonts/opentype/noto/NotoSansCJK-Regular.ttc",
+            "/usr/share/fonts/truetype/droid/DroidSansFallback.ttf",
+        ]
+    
+    # 查找第一个存在的字体文件
+    for font_path in font_locations:
+        if os.path.exists(font_path):
+            logger.info(f"在系统中找到中文字体: {font_path}")
+            return font_path
+    
+    logger.warning("在系统中未找到任何中文字体")
+    return None
+
+def install_font(target_dir=None):
+    """
+    安装中文字体
+    
+    Args:
+        target_dir: 目标目录，如果不提供则使用默认目录
+        
+    Returns:
+        str: 安装后的字体文件路径，如果失败则返回None
+    """
+    # 如果没有指定目标目录，使用默认位置
+    if target_dir is None:
+        # 获取当前脚本所在目录
+        current_dir = os.path.dirname(os.path.abspath(__file__))
+        target_dir = os.path.join(current_dir, "fonts")
+    
+    # 确保目标目录存在
+    os.makedirs(target_dir, exist_ok=True)
+    
+    # 目标字体文件路径
+    target_font = os.path.join(target_dir, "simhei.ttf")
+    
+    # 如果目标字体已存在，直接返回
+    if os.path.exists(target_font):
+        logger.info(f"中文字体已存在: {target_font}")
+        return target_font
+    
+    # 尝试从系统中复制字体
+    system_font = find_system_fonts()
+    if system_font:
+        try:
+            shutil.copy2(system_font, target_font)
+            logger.info(f"已从系统复制字体: {system_font} -> {target_font}")
+            return target_font
+        except Exception as e:
+            logger.error(f"复制系统字体失败: {str(e)}")
+    
+    # 如果系统中找不到字体，尝试下载
+    logger.info("尝试下载中文字体...")
+    downloaded_font = download_wqy_font()
+    if downloaded_font:
+        try:
+            shutil.copy2(downloaded_font, target_font)
+            logger.info(f"已安装下载的字体: {downloaded_font} -> {target_font}")
+            return target_font
+        except Exception as e:
+            logger.error(f"复制下载的字体失败: {str(e)}")
+    
+    logger.error("安装中文字体失败")
+    return None
+
+def main():
+    """
+    主函数
+    """
+    print("中文字体安装工具")
+    print("----------------")
+    
+    target_dir = None
+    if len(sys.argv) > 1:
+        target_dir = sys.argv[1]
+        print(f"使用指定的目标目录: {target_dir}")
+    
+    result = install_font(target_dir)
+    
+    if result:
+        print(f"\n✓ 成功安装中文字体: {result}")
+        print("\n现在可以正常生成PDF报告了。")
+    else:
+        print("\n✗ 安装中文字体失败")
+        print("\n请手动将中文字体(.ttf或.ttc文件)复制到以下目录:")
+        if target_dir:
+            print(f"  {target_dir}")
+        else:
+            print(f"  {os.path.join(os.path.dirname(os.path.abspath(__file__)), 'fonts')}")
+        print("\n并将其重命名为'simhei.ttf'")
+
+if __name__ == "__main__":
+    main() 

src/fundamentals_llm/text_processor.py

@@ -0,0 +1,172 @@
+import re
+import logging
+from typing import Optional, List, Dict, Any
+
+logger = logging.getLogger(__name__)
+
+class TextProcessor:
+    """大模型文本处理工具类"""
+    
+    @staticmethod
+    def clean_model_output(output: str) -> str:
+        """清理模型输出文本
+        
+        Args:
+            output: 模型输出的原始文本
+            
+        Returns:
+            str: 清理后的文本
+        """
+        try:
+            # 移除引用标记
+            output = re.sub(r'\[[^\]]*\]', '', output)
+            
+            # 移除多余的空白字符
+            output = re.sub(r'\s+', ' ', output)
+            
+            # 移除首尾的空白字符
+            output = output.strip()
+            
+            # 移除可能存在的HTML标签
+            output = re.sub(r'<[^>]+>', '', output)
+            
+            # 移除可能存在的Markdown格式
+            output = re.sub(r'[*_`~]', '', output)
+            
+            return output
+            
+        except Exception as e:
+            logger.error(f"清理模型输出失败: {str(e)}")
+            return output
+
+    @staticmethod
+    def clean_thought_process(output: str) -> str:
+        """清理模型输出，移除推理过程，只保留最终结果
+
+        Args:
+            output: 模型原始输出文本
+
+        Returns:
+            str: 清理后的输出文本
+        """
+        try:
+            # 找到</think>标签的位置
+            think_end = output.find('</think>')
+            if think_end != -1:
+                # 移除</think>标签及其之前的所有内容
+                output = output[think_end + len('</think>'):]
+
+            # 处理可能存在的空行
+            lines = output.split('\n')
+            cleaned_lines = []
+            for line in lines:
+                line = line.strip()
+                if line:  # 只保留非空行
+                    cleaned_lines.append(line)
+
+            # 重新组合文本
+            output = '\n'.join(cleaned_lines)
+
+            return output.strip()
+
+        except Exception as e:
+            logger.error(f"清理模型输出失败: {str(e)}")
+            return output.strip()
+
+    @staticmethod
+    def extract_numeric_value_from_response(response: str) -> str:
+        """从模型响应中提取数值
+        
+        Args:
+            response: 模型响应的文本
+            
+        Returns:
+            str: 提取出的数值
+        """
+        try:
+            # 清理响应文本
+            cleaned_response = TextProcessor.clean_model_output(response)
+            
+            # 尝试提取数值
+            numeric_patterns = [
+                r'[-+]?\d*\.\d+',  # 浮点数
+                r'[-+]?\d+',       # 整数
+                r'[-+]?\d+\.\d*'   # 可能带小数点的数
+            ]
+            
+            for pattern in numeric_patterns:
+                matches = re.findall(pattern, cleaned_response)
+                if matches:
+                    # 返回第一个匹配的数值
+                    return matches[0]
+            
+            # 如果没有找到数值，返回原始响应
+            return cleaned_response
+            
+        except Exception as e:
+            logger.error(f"提取数值失败: {str(e)}")
+            return response
+    
+    @staticmethod
+    def extract_list_from_response(response: str) -> List[str]:
+        """从模型响应中提取列表项
+        
+        Args:
+            response: 模型响应的文本
+            
+        Returns:
+            List[str]: 提取出的列表项
+        """
+        try:
+            # 清理响应文本
+            cleaned_response = TextProcessor.clean_model_output(response)
+            
+            # 尝试匹配列表项
+            # 匹配数字编号的列表项
+            numbered_items = re.findall(r'\d+\.\s*([^\n]+)', cleaned_response)
+            if numbered_items:
+                return [item.strip() for item in numbered_items]
+            
+            # 匹配带符号的列表项
+            bullet_items = re.findall(r'[-*•]\s*([^\n]+)', cleaned_response)
+            if bullet_items:
+                return [item.strip() for item in bullet_items]
+            
+            # 如果没有找到列表项，返回空列表
+            return []
+            
+        except Exception as e:
+            logger.error(f"提取列表项失败: {str(e)}")
+            return []
+    
+    @staticmethod
+    def extract_key_value_pairs(response: str) -> Dict[str, str]:
+        """从模型响应中提取键值对
+        
+        Args:
+            response: 模型响应的文本
+            
+        Returns:
+            Dict[str, str]: 提取出的键值对
+        """
+        try:
+            # 清理响应文本
+            cleaned_response = TextProcessor.clean_model_output(response)
+            
+            # 尝试匹配键值对
+            # 匹配冒号分隔的键值对
+            pairs = re.findall(r'([^:]+):\s*([^\n]+)', cleaned_response)
+            if pairs:
+                return {key.strip(): value.strip() for key, value in pairs}
+            
+            # 匹配等号分隔的键值对
+            pairs = re.findall(r'([^=]+)=\s*([^\n]+)', cleaned_response)
+            if pairs:
+                return {key.strip(): value.strip() for key, value in pairs}
+            
+            # 如果没有找到键值对，返回空字典
+            return {}
+            
+        except Exception as e:
+            logger.error(f"提取键值对失败: {str(e)}")
+            return {} 

src/scripts/config.py

@@ -66,12 +66,22 @@ MODEL_CONFIGS = {
             "doubao": "doubao-1-5-pro-32k-250115"
         }
     },
+    # 谷歌Gemini
+    "Gemini": {
+        "base_url": "https://generativelanguage.googleapis.com/v1beta/openai/",
+        "api_key": "AIzaSyAVE8yTaPtN-TxCCHTc9Jb-aCV-Xo1EFuU",
+        "models": {
+            "offline_model": "gemini-2.0-flash"
+        }
+    },
     # 天链苹果
     "tl_private": {
-        "base_url": "http://192.168.32.118:1234/v1/",
+        "base_url": "http://192.168.16.174:1234/v1/",
         "api_key": "none",
         "models": {
-            "ds-v1": "mlx-community/DeepSeek-R1-4bit"
+            "glm-z1": "glm-z1-rumination-32b-0414",
+            "glm-4": "glm-4-32b-0414-abliterated",
+            "ds_v1": "mlx-community/DeepSeek-R1-4bit",
         }
     },
     # 天链-千问
@@ -79,7 +89,8 @@ MODEL_CONFIGS = {
         "base_url": "http://192.168.16.178:11434/v1",
         "api_key": "sk-WaVRJKkyhrFlH4ZV35B9Aa61759b400c9cA002D00f3f1019",
         "models": {
-            "qwq": "qwq:32b"
+            "qwq": "qwq:32b",
+            "GLM": "hf-mirror.com/Cobra4687/GLM-4-32B-0414-abliterated-Q4_K_M-GGUF:Q4_K_M"
         }
     },
     # Deepseek配置

src/stock_analysis_v2 copy.py

@@ -0,0 +1,986 @@
+import pandas as pd
+import numpy as np
+from sqlalchemy import create_engine
+from datetime import datetime, timedelta
+from tqdm import tqdm
+import matplotlib.pyplot as plt
+import os
+
+# v2版本只做表格
+
+# 添加中文字体支持
+plt.rcParams['font.sans-serif'] = ['SimHei']  # 用来正常显示中文标签
+plt.rcParams['axes.unicode_minus'] = False  # 用来正常显示负号
+
+class StockAnalyzer:
+    def __init__(self, db_connection_string):
+        """
+        Initialize the stock analyzer
+        :param db_connection_string: Database connection string
+        """
+        self.engine = create_engine(db_connection_string)
+        self.initial_capital = 300000  # 30万本金
+        self.total_position = 1000000  # 100万建仓规模
+        self.borrowed_capital = 700000  # 70万借入资金
+        self.annual_interest_rate = 0.05  # 年化5%利息
+        self.daily_interest_rate = self.annual_interest_rate / 365
+        self.commission_rate = 0.05  # 5%手续费
+        self.min_holding_days = 7  # 最少持仓天数
+
+    def get_stock_list(self):
+        """获取所有股票列表"""
+        # query = "SELECT DISTINCT gp_code as symbol FROM gp_code_all where mark1 = '1'"
+        # return pd.read_sql(query, self.engine)['symbol'].tolist()
+        return ['SH600522',
+                'SZ002340',
+                'SZ000733',
+                'SH601615',
+                'SH600157',
+                'SH688005',
+                'SH600903',
+                'SH600956',
+                'SH601187',
+                'SH603983']  # 你可以在这里添加更多股票代码
+
+    def get_stock_data(self, symbol, start_date, end_date, include_history=False):
+        """获取指定股票在时间范围内的数据"""
+        try:
+            if include_history:
+                # 如果需要历史数据，获取start_date之前240天的数据
+                query = f"""
+                WITH history_data AS (
+                    SELECT 
+                        symbol,
+                        timestamp,
+                        CAST(close as DECIMAL(10,2)) as close
+                    FROM gp_day_data 
+                    WHERE symbol = '{symbol}' 
+                    AND timestamp < '{start_date}'
+                    ORDER BY timestamp DESC
+                    LIMIT 240
+                )
+                SELECT 
+                    symbol,
+                    timestamp,
+                    CAST(open as DECIMAL(10,2)) as open,
+                    CAST(high as DECIMAL(10,2)) as high,
+                    CAST(low as DECIMAL(10,2)) as low,
+                    CAST(close as DECIMAL(10,2)) as close,
+                    CAST(volume as DECIMAL(20,0)) as volume
+                FROM gp_day_data 
+                WHERE symbol = '{symbol}' 
+                AND timestamp BETWEEN '{start_date}' AND '{end_date}'
+                UNION ALL
+                SELECT 
+                    symbol,
+                    timestamp,
+                    NULL as open,
+                    NULL as high,
+                    NULL as low,
+                    close,
+                    NULL as volume
+                FROM history_data
+                ORDER BY timestamp
+                """
+            else:
+                query = f"""
+                SELECT 
+                    symbol,
+                    timestamp,
+                    CAST(open as DECIMAL(10,2)) as open,
+                    CAST(high as DECIMAL(10,2)) as high,
+                    CAST(low as DECIMAL(10,2)) as low,
+                    CAST(close as DECIMAL(10,2)) as close,
+                    CAST(volume as DECIMAL(20,0)) as volume
+                FROM gp_day_data 
+                WHERE symbol = '{symbol}' 
+                AND timestamp BETWEEN '{start_date}' AND '{end_date}'
+                ORDER BY timestamp
+                """
+            
+            # 使用chunksize分批读取数据
+            chunks = []
+            for chunk in pd.read_sql(query, self.engine, chunksize=1000):
+                chunks.append(chunk)
+            df = pd.concat(chunks, ignore_index=True)
+            
+            # 确保数值列的类型正确
+            numeric_columns = ['open', 'high', 'low', 'close', 'volume']
+            for col in numeric_columns:
+                if col in df.columns:
+                    df[col] = pd.to_numeric(df[col], errors='coerce')
+            
+            # 删除任何包含 NaN 的行，但只考虑分析期间需要的列
+            if include_history:
+                historical_mask = (df['timestamp'] < start_date) & df['close'].notna()
+                analysis_mask = (df['timestamp'] >= start_date) & df[['open', 'high', 'low', 'close', 'volume']].notna().all(axis=1)
+                df = df[historical_mask | analysis_mask]
+            else:
+                df = df.dropna()
+            
+            return df
+            
+        except Exception as e:
+            print(f"获取股票 {symbol} 数据时出错: {str(e)}")
+            return pd.DataFrame()
+
+    def calculate_trade_signals(self, df, take_profit_pct, stop_loss_pct, pullback_entry_pct, start_date):
+        """计算交易信号和收益"""
+        """计算交易信号和收益"""
+        # 检查历史数据是否足够
+        historical_data = df[df['timestamp'] < start_date]
+        if len(historical_data) < 240:
+            # print(f"历史数据不足240天，跳过分析")
+            return pd.DataFrame()
+
+        # 计算240日均线基准价格（使用历史数据）
+        ma240_base_price = historical_data['close'].tail(240).mean()
+        # print(f"240日均线基准价格: {ma240_base_price:.2f}")
+
+        # 只使用分析期间的数据进行交易信号计算
+        analysis_df = df[df['timestamp'] >= start_date].copy()
+        if len(analysis_df) < 10:  # 确保分析期间有足够的数据
+            # print(f"分析期间数据不足10天，跳过分析")
+            return pd.DataFrame()
+
+        results = []
+        positions = []  # 记录所有持仓，每个元素是一个字典，包含入场价格和日期
+        max_positions = 5  # 最大建仓次数
+        cumulative_profit = 0  # 总累计收益
+        trades_count = 0
+        max_holding_period = timedelta(days=180)  # 最长持仓6个月
+        max_loss_amount = -300000  # 最大亏损金额为30万
+        
+        # 添加统计变量
+        total_holding_days = 0  # 总持仓天数
+        max_capital_usage = 0  # 最大资金占用
+        total_trades = 0  # 总交易次数
+        profitable_trades = 0  # 盈利交易次数
+        loss_trades = 0  # 亏损交易次数
+
+        # print(f"\n{'='*50}")
+        # print(f"开始分析股票: {df.iloc[0]['symbol']}")
+        # print(f"{'='*50}")
+
+        # 记录前一天的收盘价
+        prev_close = None
+
+        for index, row in analysis_df.iterrows():
+            current_price = (float(row['high']) + float(row['low'])) / 2
+            day_result = {
+                'timestamp': row['timestamp'],
+                'profit': cumulative_profit,  # 记录当前的累计收益
+                'position': False,
+                'action': 'hold'
+            }
+
+            # 更新所有持仓的收益
+            positions_to_remove = []
+
+            for pos in positions:
+                pos_days_held = (row['timestamp'] - pos['entry_date']).days
+                pos_size = self.total_position / pos['entry_price']
+                
+                # 检查是否需要平仓
+                should_close = False
+                close_type = None
+                close_price = current_price
+
+                if pos_days_held >= self.min_holding_days:
+                    # 检查止盈
+                    if float(row['high']) >= pos['entry_price'] * (1 + take_profit_pct):
+                        should_close = True
+                        close_type = 'sell_profit'
+                        close_price = pos['entry_price'] * (1 + take_profit_pct)
+                    # 检查最大亏损
+                    elif (close_price - pos['entry_price']) * pos_size <= max_loss_amount:
+                        should_close = True
+                        close_type = 'sell_loss'
+                    # 检查持仓时间
+                    elif (row['timestamp'] - pos['entry_date']) >= max_holding_period:
+                        should_close = True
+                        close_type = 'time_limit'
+
+                if should_close:
+                    profit = (close_price - pos['entry_price']) * pos_size
+                    commission = profit * self.commission_rate if profit > 0 else 0
+                    interest_cost = self.borrowed_capital * self.daily_interest_rate * pos_days_held
+                    net_profit = profit - commission - interest_cost
+                    
+                    # 更新统计数据
+                    total_holding_days += pos_days_held
+                    total_trades += 1
+                    if net_profit > 0:
+                        profitable_trades += 1
+                    else:
+                        loss_trades += 1
+                    
+                    cumulative_profit += net_profit
+                    positions_to_remove.append(pos)
+                    day_result['action'] = close_type
+                    day_result['profit'] = cumulative_profit
+                    
+                    # print(f"\n>>> {'止盈平仓' if close_type == 'sell_profit' else '止损平仓' if close_type == 'sell_loss' else '到期平仓'}")
+                    # print(f"持仓天数: {pos_days_held}")
+                    # print(f"持仓数量: {pos_size:.0f}股")
+                    # print(f"平仓价格: {close_price:.2f}")
+                    # print(f"{'盈利' if profit > 0 else '亏损'}: {profit:,.2f}")
+                    # print(f"手续费: {commission:,.2f}")
+                    # print(f"利息成本: {interest_cost:,.2f}")
+                    # print(f"净{'盈利' if net_profit > 0 else '亏损'}: {net_profit:,.2f}")
+
+            # 移除已平仓的持仓
+            for pos in positions_to_remove:
+                positions.remove(pos)
+
+            # 检查是否可以建仓
+            if len(positions) < max_positions and prev_close is not None:
+                # 计算当日跌幅
+                daily_drop = (prev_close - float(row['low'])) / prev_close
+                
+                # 检查价格是否在均线区间内
+                price_in_range = (
+                    current_price >= ma240_base_price * 0.7 and 
+                    current_price <= ma240_base_price * 1.3
+                )
+                
+                # 如果跌幅超过5%且价格在均线区间内
+                if daily_drop >= 0.05 and price_in_range:
+                    positions.append({
+                        'entry_price': current_price,
+                        'entry_date': row['timestamp']
+                    })
+                    trades_count += 1
+                    # print(f"\n>>> 建仓信号 #{trades_count}")
+                    # print(f"日期: {row['timestamp'].strftime('%Y-%m-%d')}")
+                    # print(f"建仓价格: {current_price:.2f}")
+                    # print(f"建仓数量: {self.total_position/current_price:.0f}股")
+                    # print(f"当日跌幅: {daily_drop*100:.2f}%")
+                    # print(f"距离均线: {((current_price/ma240_base_price)-1)*100:.2f}%")
+
+            # 更新前一天收盘价
+            prev_close = float(row['close'])
+
+            # 更新日结果
+            day_result['position'] = len(positions) > 0
+            results.append(day_result)
+
+            # 计算当前资金占用
+            current_capital_usage = sum(self.total_position for pos in positions)
+            max_capital_usage = max(max_capital_usage, current_capital_usage)
+
+        # 在最后一天强制平仓所有持仓
+        if positions:
+            final_price = (float(analysis_df.iloc[-1]['high']) + float(analysis_df.iloc[-1]['low'])) / 2
+            final_total_profit = 0
+            
+            for pos in positions:
+                position_size = self.total_position / pos['entry_price']
+                days_held = (analysis_df.iloc[-1]['timestamp'] - pos['entry_date']).days
+                final_profit = (final_price - pos['entry_price']) * position_size
+                commission = final_profit * self.commission_rate if final_profit > 0 else 0
+                interest_cost = self.borrowed_capital * self.daily_interest_rate * days_held
+                net_profit = final_profit - commission - interest_cost
+                
+                # 更新统计数据
+                total_holding_days += days_held
+                total_trades += 1
+                if net_profit > 0:
+                    profitable_trades += 1
+                else:
+                    loss_trades += 1
+                
+                final_total_profit += net_profit
+                
+                # print(f"\n>>> 到期强制平仓")
+                # print(f"持仓天数: {days_held}")
+                # print(f"持仓数量: {position_size:.0f}股")
+                # print(f"平仓价格: {final_price:.2f}")
+                # print(f"毛利润: {final_profit:,.2f}")
+                # print(f"手续费: {commission:,.2f}")
+                # print(f"利息成本: {interest_cost:,.2f}")
+                # print(f"净利润: {net_profit:,.2f}")
+            
+            results[-1]['action'] = 'final_sell'
+            cumulative_profit += final_total_profit  # 更新最终的累计收益
+            results[-1]['profit'] = cumulative_profit  # 更新最后一天的累计收益
+
+        # 计算统计数据
+        avg_holding_days = total_holding_days / total_trades if total_trades > 0 else 0
+        win_rate = profitable_trades / total_trades * 100 if total_trades > 0 else 0
+
+        # print(f"\n{'='*50}")
+        # print(f"交易统计")
+        # print(f"总交易次数: {trades_count}")
+        # print(f"累计收益: {cumulative_profit:,.2f}")
+        # print(f"最大资金占用: {max_capital_usage:,.2f}")
+        # print(f"平均持仓天数: {avg_holding_days:.1f}")
+        # print(f"胜率: {win_rate:.1f}%")
+        # print(f"盈利交易: {profitable_trades}次")
+        # print(f"亏损交易: {loss_trades}次")
+        # print(f"{'='*50}\n")
+
+        # 将统计数据添加到结果中
+        if len(results) > 0:
+            results[-1]['stats'] = {
+                'total_trades': total_trades,
+                'profitable_trades': profitable_trades,
+                'loss_trades': loss_trades,
+                'win_rate': win_rate,
+                'avg_holding_days': avg_holding_days,
+                'max_capital_usage': max_capital_usage,
+                'final_profit': cumulative_profit
+            }
+
+        return pd.DataFrame(results)
+
+    def analyze_stock(self, symbol, start_date, end_date, take_profit_pct, stop_loss_pct, pullback_entry_pct):
+        """分析单个股票"""
+        df = self.get_stock_data(symbol, start_date, end_date, include_history=True)
+        
+        if len(df) < 240:  # 确保总数据量足够
+            print(f"股票 {symbol} 总数据不足240天，跳过分析")
+            return None
+            
+        return self.calculate_trade_signals(
+            df,
+            take_profit_pct=take_profit_pct,
+            stop_loss_pct=stop_loss_pct,
+            pullback_entry_pct=pullback_entry_pct,
+            start_date=start_date
+        )
+
+    def analyze_all_stocks(self, start_date, end_date, take_profit_pct, stop_loss_pct, pullback_entry_pct):
+        """分析所有股票"""
+        stocks = self.get_stock_list()
+        all_results = {}
+        analysis_summary = {
+            'processed': [],
+            'skipped': [],
+            'error': []
+        }
+        
+        print(f"\n开始分析，共 {len(stocks)} 支股票")
+        
+        for symbol in tqdm(stocks, desc="Analyzing stocks"):
+            try:
+                # 获取股票数据
+                df = self.get_stock_data(symbol, start_date, end_date, include_history=True)
+                
+                if df.empty:
+                    print(f"\n股票 {symbol} 数据为空，跳过分析")
+                    analysis_summary['skipped'].append({
+                        'symbol': symbol,
+                        'reason': 'empty_data',
+                        'take_profit_pct': take_profit_pct
+                    })
+                    continue
+                
+                # 检查历史数据
+                historical_data = df[df['timestamp'] < start_date]
+                if len(historical_data) < 240:
+                    # print(f"\n股票 {symbol} 历史数据不足240天({len(historical_data)}天)，跳过分析")
+                    analysis_summary['skipped'].append({
+                        'symbol': symbol,
+                        'reason': f'insufficient_history_{len(historical_data)}days',
+                        'take_profit_pct': take_profit_pct
+                    })
+                    continue
+                
+                # 检查分析期数据
+                analysis_data = df[df['timestamp'] >= start_date]
+                if len(analysis_data) < 10:
+                    print(f"\n股票 {symbol} 分析期数据不足10天({len(analysis_data)}天)，跳过分析")
+                    analysis_summary['skipped'].append({
+                        'symbol': symbol,
+                        'reason': f'insufficient_analysis_data_{len(analysis_data)}days',
+                        'take_profit_pct': take_profit_pct
+                    })
+                    continue
+                
+                # 计算交易信号
+                result = self.calculate_trade_signals(
+                    df,
+                    take_profit_pct,
+                    stop_loss_pct,
+                    pullback_entry_pct,
+                    start_date
+                )
+                
+                if result is not None and not result.empty:
+                    # 检查是否有交易记录
+                    has_trades = any(action in ['sell_profit', 'sell_loss', 'time_limit', 'final_sell'] 
+                                   for action in result['action'] if pd.notna(action))
+                    
+                    if has_trades:
+                        all_results[symbol] = result
+                        analysis_summary['processed'].append({
+                            'symbol': symbol,
+                            'take_profit_pct': take_profit_pct,
+                            'trades': len([x for x in result['action'] if pd.notna(x) and x != 'hold']),
+                            'profit': result['profit'].iloc[-1]
+                        })
+                    else:
+                        # print(f"\n股票 {symbol} 没有产生有效的交易信号")
+                        analysis_summary['skipped'].append({
+                            'symbol': symbol,
+                            'reason': 'no_valid_signals',
+                            'take_profit_pct': take_profit_pct
+                        })
+                
+            except Exception as e:
+                print(f"\n处理股票 {symbol} 时出错: {str(e)}")
+                analysis_summary['error'].append({
+                    'symbol': symbol,
+                    'error': str(e),
+                    'take_profit_pct': take_profit_pct
+                })
+                continue
+        
+        # 打印分析总结
+        print(f"\n分析完成:")
+        print(f"成功分析: {len(analysis_summary['processed'])} 支股票")
+        print(f"跳过分析: {len(analysis_summary['skipped'])} 支股票")
+        print(f"出错股票: {len(analysis_summary['error'])} 支股票")
+        
+        # 保存分析总结
+        if not os.path.exists('results/analysis_summary'):
+            os.makedirs('results/analysis_summary')
+        
+        # 保存处理成功的股票信息
+        if analysis_summary['processed']:
+            pd.DataFrame(analysis_summary['processed']).to_csv(
+                f'results/analysis_summary/processed_stocks_{take_profit_pct*100:.1f}.csv',
+                index=False
+            )
+        
+        # 保存跳过的股票信息
+        if analysis_summary['skipped']:
+            pd.DataFrame(analysis_summary['skipped']).to_csv(
+                f'results/analysis_summary/skipped_stocks_{take_profit_pct*100:.1f}.csv',
+                index=False
+            )
+        
+        # 保存出错的股票信息
+        if analysis_summary['error']:
+            pd.DataFrame(analysis_summary['error']).to_csv(
+                f'results/analysis_summary/error_stocks_{take_profit_pct*100:.1f}.csv',
+                index=False
+            )
+        
+        return all_results
+
+    def plot_results(self, results, symbol):
+        """绘制单个股票的收益走势图"""
+        plt.figure(figsize=(12, 6))
+        plt.plot(results['timestamp'], results['profit'].cumsum(), label='Cumulative Profit')
+        plt.title(f'Stock {symbol} Trading Results')
+        plt.xlabel('Date')
+        plt.ylabel('Profit (CNY)')
+        plt.legend()
+        plt.grid(True)
+        plt.xticks(rotation=45)
+        plt.tight_layout()
+        return plt
+
+    def plot_all_stocks(self, all_results):
+        """绘制所有股票的收益走势图"""
+        plt.figure(figsize=(15, 8))
+        
+        # 记录所有股票的累计收益
+        total_profits = {}
+        max_profit = float('-inf')
+        min_profit = float('inf')
+        
+        # 绘制每支股票的曲线
+        for symbol, results in all_results.items():
+            # 直接使用已经计算好的累计收益
+            profits = results['profit']
+            plt.plot(results['timestamp'], profits, label=symbol, alpha=0.5, linewidth=1)
+            
+            # 更新最大最小收益
+            max_profit = max(max_profit, profits.max())
+            min_profit = min(min_profit, profits.min())
+            
+            # 记录总收益（使用最后一个累计收益值）
+            total_profits[symbol] = profits.iloc[-1]
+        
+        # 计算所有股票的中位数收益曲线
+        # 首先找到共同的日期范围
+        all_dates = set()
+        for results in all_results.values():
+            all_dates.update(results['timestamp'])
+        all_dates = sorted(list(all_dates))
+        
+        # 创建一个包含所有日期的DataFrame
+        profits_df = pd.DataFrame(index=all_dates)
+        
+        # 对每支股票，填充所有日期的收益
+        for symbol, results in all_results.items():
+            daily_profits = pd.Series(index=all_dates, data=np.nan)
+            # 直接使用已经计算好的累计收益
+            for date, profit in zip(results['timestamp'], results['profit']):
+                daily_profits[date] = profit
+            # 对于没有交易的日期，使用最后一个累计收益填充
+            daily_profits.fillna(method='ffill', inplace=True)
+            daily_profits.fillna(0, inplace=True)  # 对开始之前的日期填充0
+            profits_df[symbol] = daily_profits
+        
+        # 计算并绘制中位数收益曲线
+        median_line = profits_df.median(axis=1)
+        plt.plot(all_dates, median_line, 'r-', label='Median', linewidth=2)
+        
+        plt.title('All Stocks Trading Results')
+        plt.xlabel('Date')
+        plt.ylabel('Profit (CNY)')
+        plt.grid(True)
+        plt.xticks(rotation=45)
+        
+        # 添加利润区间标注
+        plt.text(0.02, 0.98, 
+                f'Profit Range:\nMax: {max_profit:,.0f}\nMin: {min_profit:,.0f}\nMedian: {median_line.iloc[-1]:,.0f}', 
+                transform=plt.gca().transAxes, 
+                bbox=dict(facecolor='white', alpha=0.8))
+        
+        # 计算所有股票的统计数据
+        total_stats = {
+            'total_trades': 0,
+            'profitable_trades': 0,
+            'loss_trades': 0,
+            'total_holding_days': 0,
+            'max_capital_usage': 0,
+            'total_profit': 0,
+            'profitable_stocks': 0,
+            'loss_stocks': 0,
+            'total_capital_usage': 0,  # 添加总资金占用
+            'total_win_rate': 0,       # 添加总胜率
+        }
+
+        # 获取前5名和后5名股票
+        top_5 = sorted(total_profits.items(), key=lambda x: x[1], reverse=True)[:5]
+        bottom_5 = sorted(total_profits.items(), key=lambda x: x[1])[:5]
+        
+        # 添加排名信息
+        rank_text = "Top 5 Stocks:\n"
+        for symbol, profit in top_5:
+            rank_text += f"{symbol}: {profit:,.0f}\n"
+        rank_text += "\nBottom 5 Stocks:\n"
+        for symbol, profit in bottom_5:
+            rank_text += f"{symbol}: {profit:,.0f}\n"
+            
+        plt.text(1.02, 0.98, rank_text, transform=plt.gca().transAxes, 
+                bbox=dict(facecolor='white', alpha=0.8), verticalalignment='top')
+
+        # 记录每支股票的统计数据用于计算平均值
+        stock_stats = []
+
+        for symbol, results in all_results.items():
+            if 'stats' in results.iloc[-1]:
+                stats = results.iloc[-1]['stats']
+                stock_stats.append(stats)  # 记录每支股票的统计数据
+                
+                total_stats['total_trades'] += stats['total_trades']
+                total_stats['profitable_trades'] += stats['profitable_trades']
+                total_stats['loss_trades'] += stats['loss_trades']
+                total_stats['total_holding_days'] += stats['avg_holding_days'] * stats['total_trades']
+                total_stats['max_capital_usage'] = max(total_stats['max_capital_usage'], stats['max_capital_usage'])
+                total_stats['total_capital_usage'] += stats['max_capital_usage']
+                total_stats['total_profit'] += stats['final_profit']
+                total_stats['total_win_rate'] += stats['win_rate']
+                if stats['final_profit'] > 0:
+                    total_stats['profitable_stocks'] += 1
+                else:
+                    total_stats['loss_stocks'] += 1
+
+        # 计算总体统计
+        total_stocks = total_stats['profitable_stocks'] + total_stats['loss_stocks']
+        avg_holding_days = total_stats['total_holding_days'] / total_stats['total_trades'] if total_stats['total_trades'] > 0 else 0
+        win_rate = total_stats['profitable_trades'] / total_stats['total_trades'] * 100 if total_stats['total_trades'] > 0 else 0
+        stock_win_rate = total_stats['profitable_stocks'] / total_stocks * 100 if total_stocks > 0 else 0
+        
+        # 计算平均值统计
+        avg_trades_per_stock = total_stats['total_trades'] / total_stocks if total_stocks > 0 else 0
+        avg_profit_per_stock = total_stats['total_profit'] / total_stocks if total_stocks > 0 else 0
+        avg_capital_usage = total_stats['total_capital_usage'] / total_stocks if total_stocks > 0 else 0
+        avg_win_rate = total_stats['total_win_rate'] / total_stocks if total_stocks > 0 else 0
+
+        # 添加总体统计信息
+        stats_text = (
+            f"总体统计:\n"
+            f"总交易次数: {total_stats['total_trades']}\n"
+            # f"最大资金占用: {total_stats['max_capital_usage']:,.0f}\n"
+            f"平均持仓天数: {avg_holding_days:.1f}\n"
+            f"交易胜率: {win_rate:.1f}%\n"
+            f"股票胜率: {stock_win_rate:.1f}%\n"
+            f"盈利股票数: {total_stats['profitable_stocks']}\n"
+            f"亏损股票数: {total_stats['loss_stocks']}\n"
+            f"\n平均值统计:\n"
+            f"平均交易次数: {avg_trades_per_stock:.1f}\n"
+            f"平均收益: {avg_profit_per_stock:,.0f}\n"
+            f"平均资金占用: {avg_capital_usage:,.0f}\n"
+            f"平均持仓天数: {avg_holding_days:.1f}\n"
+            f"平均胜率: {avg_win_rate:.1f}%"
+        )
+        
+        plt.text(0.02, 0.6, stats_text,
+                transform=plt.gca().transAxes,
+                bbox=dict(facecolor='white', alpha=0.8, edgecolor='gray'),
+                verticalalignment='top',
+                fontsize=10)
+        
+        # 获取前5名和后5名股票
+        top_5 = sorted(total_profits.items(), key=lambda x: x[1], reverse=True)[:5]
+        bottom_5 = sorted(total_profits.items(), key=lambda x: x[1])[:5]
+        
+        # 添加排名信息
+        rank_text = "Top 5 Stocks:\n"
+        for symbol, profit in top_5:
+            rank_text += f"{symbol}: {profit:,.0f}\n"
+        rank_text += "\nBottom 5 Stocks:\n"
+        for symbol, profit in bottom_5:
+            rank_text += f"{symbol}: {profit:,.0f}\n"
+            
+        plt.text(1.02, 0.98, rank_text, transform=plt.gca().transAxes, 
+                bbox=dict(facecolor='white', alpha=0.8), verticalalignment='top')
+        
+        plt.tight_layout()
+        return plt, total_profits
+
+    def plot_profit_distribution(self, total_profits):
+        """绘制所有股票的收益分布小提琴图"""
+        plt.figure(figsize=(12, 8))
+        
+        # 将收益数据转换为数组，确保使用最终累计收益
+        profits = np.array([profit for profit in total_profits.values()])
+        
+        # 计算统计数据
+        mean_profit = np.mean(profits)
+        median_profit = np.median(profits)
+        max_profit = np.max(profits)
+        min_profit = np.min(profits)
+        std_profit = np.std(profits)
+        
+        # 绘制小提琴图
+        violin_parts = plt.violinplot(profits, positions=[0], showmeans=True, showmedians=True)
+        
+        # 设置小提琴图的颜色
+        violin_parts['bodies'][0].set_facecolor('lightblue')
+        violin_parts['bodies'][0].set_alpha(0.7)
+        violin_parts['cmeans'].set_color('red')
+        violin_parts['cmedians'].set_color('blue')
+        
+        # 添加统计信息
+        stats_text = (
+            f"统计信息:\n"
+            f"平均收益: {mean_profit:,.0f}\n"
+            f"中位数收益: {median_profit:,.0f}\n"
+            f"最大收益: {max_profit:,.0f}\n"
+            f"最小收益: {min_profit:,.0f}\n"
+            f"标准差: {std_profit:,.0f}"
+        )
+        
+        plt.text(0.65, 0.95, stats_text,
+                transform=plt.gca().transAxes,
+                bbox=dict(facecolor='white', alpha=0.8, edgecolor='gray'),
+                verticalalignment='top',
+                fontsize=10)
+        
+        # 设置图表样式
+        plt.title('股票收益分布图', fontsize=14, pad=20)
+        plt.ylabel('收益 (元)', fontsize=12)
+        plt.xticks([0], ['所有股票'], fontsize=10)
+        plt.grid(True, axis='y', alpha=0.3)
+        
+        # 添加零线
+        plt.axhline(y=0, color='r', linestyle='--', alpha=0.3)
+        
+        # 计算盈利和亏损的股票数量
+        profit_count = np.sum(profits > 0)
+        loss_count = np.sum(profits < 0)
+        total_count = len(profits)
+        
+        # 添加盈亏比例信息
+        ratio_text = (
+            f"盈亏比例:\n"
+            f"盈利: {profit_count}支 ({profit_count/total_count*100:.1f}%)\n"
+            f"亏损: {loss_count}支 ({loss_count/total_count*100:.1f}%)\n"
+            f"总计: {total_count}支"
+        )
+        
+        plt.text(0.65, 0.6, ratio_text,
+                transform=plt.gca().transAxes,
+                bbox=dict(facecolor='white', alpha=0.8, edgecolor='gray'),
+                verticalalignment='top',
+                fontsize=10)
+        
+        plt.tight_layout()
+        return plt
+
+    def plot_profit_matrix(self, df):
+        """绘制止盈比例分析矩阵的结果图"""
+        plt.figure(figsize=(15, 10))
+        
+        # 创建子图
+        fig, ((ax1, ax2), (ax3, ax4)) = plt.subplots(2, 2, figsize=(15, 10))
+        
+        # 1. 总收益曲线
+        ax1.plot(df.index, df['total_profit'], 'b-', linewidth=2)
+        ax1.set_title('总收益 vs 止盈比例')
+        ax1.set_xlabel('止盈比例 (%)')
+        ax1.set_ylabel('总收益 (元)')
+        ax1.grid(True)
+        
+        # 标记最优止盈比例点
+        best_profit_pct = df['total_profit'].idxmax()
+        best_profit = df['total_profit'].max()
+        ax1.plot(best_profit_pct, best_profit, 'ro')
+        ax1.annotate(f'最优: {best_profit_pct:.1f}%\n{best_profit:,.0f}元',
+                    (best_profit_pct, best_profit),
+                    xytext=(10, 10), textcoords='offset points')
+        
+        # 2. 平均每笔收益和中位数收益
+        ax2.plot(df.index, df['avg_profit_per_trade'], 'g-', linewidth=2, label='平均收益')
+        ax2.plot(df.index, df['median_profit_per_trade'], 'r--', linewidth=2, label='中位数收益')
+        ax2.set_title('每笔交易收益 vs 止盈比例')
+        ax2.set_xlabel('止盈比例 (%)')
+        ax2.set_ylabel('收益 (元)')
+        ax2.legend()
+        ax2.grid(True)
+        
+        # 3. 平均持仓天数曲线
+        ax3.plot(df.index, df['avg_holding_days'], 'r-', linewidth=2)
+        ax3.set_title('平均持仓天数 vs 止盈比例')
+        ax3.set_xlabel('止盈比例 (%)')
+        ax3.set_ylabel('持仓天数')
+        ax3.grid(True)
+        
+        # 4. 胜率对比
+        ax4.plot(df.index, df['trade_win_rate'], 'c-', linewidth=2, label='交易胜率')
+        ax4.plot(df.index, df['stock_win_rate'], 'm--', linewidth=2, label='股票胜率')
+        ax4.set_title('胜率 vs 止盈比例')
+        ax4.set_xlabel('止盈比例 (%)')
+        ax4.set_ylabel('胜率 (%)')
+        ax4.legend()
+        ax4.grid(True)
+        
+        # 调整布局
+        plt.tight_layout()
+        return plt
+
+    def analyze_profit_matrix(self, start_date, end_date, stop_loss_pct, pullback_entry_pct):
+        """分析不同止盈比例的表现矩阵"""
+        # 创建results目录和临时文件目录
+        if not os.path.exists('results/temp'):
+            os.makedirs('results/temp')
+            
+        # 从30%到5%，每次减少1%
+        profit_percentages = list(np.arange(0.30, 0.04, -0.01))
+        
+        for take_profit_pct in profit_percentages:
+            # 检查是否已经分析过这个止盈点
+            temp_file = f'results/temp/profit_{take_profit_pct*100:.1f}.csv'
+            if os.path.exists(temp_file):
+                print(f"\n止盈比例 {take_profit_pct*100:.1f}% 已分析，跳过")
+                continue
+                
+            print(f"\n分析止盈比例: {take_profit_pct*100:.1f}%")
+            
+            try:
+                # 运行分析
+                results = self.analyze_all_stocks(
+                    start_date,
+                    end_date,
+                    take_profit_pct,
+                    stop_loss_pct,
+                    pullback_entry_pct
+                )
+                
+                # 过滤掉空结果
+                valid_results = {symbol: result for symbol, result in results.items() if result is not None and not result.empty}
+                
+                if valid_results:
+                    # 计算统计数据
+                    total_stats = {
+                        'total_trades': 0,
+                        'profitable_trades': 0,
+                        'loss_trades': 0,
+                        'total_holding_days': 0,
+                        'total_profit': 0,
+                        'total_capital_usage': 0,
+                        'total_win_rate': 0,
+                        'stock_count': len(valid_results),
+                        'profitable_stocks': 0,
+                        'all_trade_profits': []
+                    }
+                    
+                    # 收集每支股票的统计数据
+                    for symbol, result in valid_results.items():
+                        if 'stats' in result.iloc[-1]:
+                            stats = result.iloc[-1]['stats']
+                            total_stats['total_trades'] += stats['total_trades']
+                            total_stats['profitable_trades'] += stats['profitable_trades']
+                            total_stats['loss_trades'] += stats['loss_trades']
+                            total_stats['total_holding_days'] += stats['avg_holding_days'] * stats['total_trades']
+                            total_stats['total_profit'] += stats['final_profit']
+                            total_stats['total_capital_usage'] += stats['max_capital_usage']
+                            total_stats['total_win_rate'] += stats['win_rate']
+                            
+                            if stats['final_profit'] > 0:
+                                total_stats['profitable_stocks'] += 1
+                                
+                            if hasattr(stats, 'trade_profits'):
+                                total_stats['all_trade_profits'].extend(stats['trade_profits'])
+                    
+                    # 计算统计数据
+                    avg_trades = total_stats['total_trades'] / total_stats['stock_count']
+                    avg_holding_days = total_stats['total_holding_days'] / total_stats['total_trades'] if total_stats['total_trades'] > 0 else 0
+                    avg_profit_per_trade = total_stats['total_profit'] / total_stats['total_trades'] if total_stats['total_trades'] > 0 else 0
+                    median_profit_per_trade = np.median(total_stats['all_trade_profits']) if total_stats['all_trade_profits'] else 0
+                    total_profit = total_stats['total_profit']
+                    trade_win_rate = (total_stats['profitable_trades'] / total_stats['total_trades'] * 100) if total_stats['total_trades'] > 0 else 0
+                    stock_win_rate = (total_stats['profitable_stocks'] / total_stats['stock_count'] * 100) if total_stats['stock_count'] > 0 else 0
+                    
+                    # 创建单个止盈点的结果DataFrame
+                    result_df = pd.DataFrame([{
+                        'take_profit_pct': take_profit_pct * 100,
+                        'total_profit': total_profit,
+                        'avg_profit_per_trade': avg_profit_per_trade,
+                        'median_profit_per_trade': median_profit_per_trade,
+                        'avg_holding_days': avg_holding_days,
+                        'trade_win_rate': trade_win_rate,
+                        'stock_win_rate': stock_win_rate,
+                        'total_trades': total_stats['total_trades'],
+                        'stock_count': total_stats['stock_count']
+                    }])
+                    
+                    # 保存单个止盈点的结果
+                    result_df.to_csv(temp_file, index=False)
+                    print(f"保存止盈点 {take_profit_pct*100:.1f}% 的分析结果")
+                    
+                # 清理内存
+                del valid_results
+                del results
+                if 'total_stats' in locals():
+                    del total_stats
+                
+            except Exception as e:
+                print(f"处理止盈点 {take_profit_pct*100:.1f}% 时出错: {str(e)}")
+                continue
+        
+        # 合并所有结果
+        print("\n合并所有分析结果...")
+        all_results = []
+        for take_profit_pct in profit_percentages:
+            temp_file = f'results/temp/profit_{take_profit_pct*100:.1f}.csv'
+            if os.path.exists(temp_file):
+                try:
+                    df = pd.read_csv(temp_file)
+                    all_results.append(df)
+                except Exception as e:
+                    print(f"读取文件 {temp_file} 时出错: {str(e)}")
+        
+        if all_results:
+            # 合并所有结果
+            df = pd.concat(all_results, ignore_index=True)
+            df = df.round({
+                'take_profit_pct': 1,
+                'total_profit': 0,
+                'avg_profit_per_trade': 0,
+                'median_profit_per_trade': 0,
+                'avg_holding_days': 1,
+                'trade_win_rate': 1,
+                'stock_win_rate': 1,
+                'total_trades': 0,
+                'stock_count': 0
+            })
+            
+            # 设置止盈比例为索引并排序
+            df.set_index('take_profit_pct', inplace=True)
+            df.sort_index(ascending=False, inplace=True)
+            
+            # 保存最终结果
+            df.to_csv('results/profit_matrix_analysis.csv')
+            
+            # 打印结果
+            print("\n止盈比例分析矩阵:")
+            print("=" * 120)
+            print(df.to_string())
+            print("=" * 120)
+            
+            try:
+                # 绘制并保存矩阵分析图表
+                plt = self.plot_profit_matrix(df)
+                plt.savefig('results/profit_matrix_analysis.png', bbox_inches='tight', dpi=300)
+                plt.close()
+            except Exception as e:
+                print(f"绘制图表时出错: {str(e)}")
+            
+            return df
+        else:
+            print("没有找到任何分析结果")
+            return pd.DataFrame()
+
+def main():
+    # 数据库连接配置
+    db_config = {
+        'host': '192.168.18.199',
+        'port': 3306,
+        'user': 'root',
+        'password': 'Chlry#$.8',
+        'database': 'db_gp_cj'
+    }
+    
+    try:
+        connection_string = f"mysql+pymysql://{db_config['user']}:{db_config['password']}@{db_config['host']}:{db_config['port']}/{db_config['database']}"
+        
+        # 创建results目录
+
+        if not os.path.exists('results'):
+            os.makedirs('results')
+        
+        # 创建分析器实例
+        analyzer = StockAnalyzer(connection_string)
+        
+        # 设置分析参数
+        start_date = '2022-05-05'
+        end_date = '2023-08-28'
+        stop_loss_pct = -0.99   # -99%止损
+        pullback_entry_pct = -0.05  # -5%回调建仓
+        
+        # 运行止盈比例分析
+        profit_matrix = analyzer.analyze_profit_matrix(
+            start_date,
+            end_date,
+            stop_loss_pct,
+            pullback_entry_pct
+        )
+        
+        if not profit_matrix.empty:
+            # 使用最优止盈比例进行完整分析
+            take_profit_pct = profit_matrix['total_profit'].idxmax() / 100
+            print(f"\n使用最优止盈比例 {take_profit_pct*100:.1f}% 运行详细分析")
+            
+            # 运行分析并只生成所需的图表
+            results = analyzer.analyze_all_stocks(
+                start_date,
+                end_date,
+                take_profit_pct,
+                stop_loss_pct,
+                pullback_entry_pct
+            )
+            
+            # 过滤掉空结果
+            valid_results = {symbol: result for symbol, result in results.items() if result is not None and not result.empty}
+            
+            if valid_results:
+                # 只生成所需的两个图表
+                plt, _ = analyzer.plot_all_stocks(valid_results)
+                plt.savefig('results/all_stocks_analysis.png', bbox_inches='tight', dpi=300)
+                plt.close()
+        
+    except Exception as e:
+        print(f"程序运行出错: {str(e)}")
+        raise
+
+if __name__ == "__main__":
+    main() 

src/tq_demo.py

@@ -0,0 +1,81 @@
+# 用前须知
+
+## xtdata提供和MiniQmt的交互接口，本质是和MiniQmt建立连接，由MiniQmt处理行情数据请求，再把结果回传返回到python层。使用的行情服务器以及能获取到的行情数据和MiniQmt是一致的，要检查数据或者切换连接时直接操作MiniQmt即可。
+
+## 对于数据获取接口，使用时需要先确保MiniQmt已有所需要的数据，如果不足可以通过补充数据接口补充，再调用数据获取接口获取。
+
+## 对于订阅接口，直接设置数据回调，数据到来时会由回调返回。订阅接收到的数据一般会保存下来，同种数据不需要再单独补充。
+
+# 代码讲解
+
+# 从本地python导入xtquant库，如果出现报错则说明安装失败
+from xtquant import xtdata
+import time
+
+# 设定一个标的列表
+code_list = ["000001.SZ"]
+# 设定获取数据的周期
+period = "1d"
+
+# 下载标的行情数据
+if 1:
+    ## 为了方便用户进行数据管理，xtquant的大部分历史数据都是以压缩形式存储在本地的
+    ## 比如行情数据，需要通过download_history_data下载，财务数据需要通过
+    ## 所以在取历史数据之前，我们需要调用数据下载接口，将数据下载到本地
+    for i in code_list:
+        xtdata.download_history_data(i, period=period, incrementally=True)  # 增量下载行情数据（开高低收,等等）到本地
+
+    xtdata.download_financial_data(code_list)  # 下载财务数据到本地
+    xtdata.download_sector_data()  # 下载板块数据到本地
+    # 更多数据的下载方式可以通过数据字典查询
+
+# 读取本地历史行情数据
+history_data = xtdata.get_market_data_ex([], code_list, period=period, count=-1)
+print(history_data)
+print("=" * 20)
+
+# 如果需要盘中的实时行情，需要向服务器进行订阅后才能获取
+# 订阅后，get_market_data函数于get_market_data_ex函数将会自动拼接本地历史行情与服务器实时行情
+
+# 向服务器订阅数据
+for i in code_list:
+    xtdata.subscribe_quote(i, period=period, count=-1)  # 设置count = -1来取到当天所有实时行情
+
+# 等待订阅完成
+time.sleep(1)
+
+# 获取订阅后的行情
+kline_data = xtdata.get_market_data_ex([], code_list, period=period)
+print(kline_data)
+
+# 获取订阅后的行情，并以固定间隔进行刷新,预期会循环打印10次
+for i in range(10):
+    # 这边做演示，就用for来循环了，实际使用中可以用while True
+    kline_data = xtdata.get_market_data_ex([], code_list, period=period)
+    print(kline_data)
+    time.sleep(3)  # 三秒后再次获取行情
+
+
+# 如果不想用固定间隔触发，可以以用订阅后的回调来执行
+# 这种模式下当订阅的callback回调函数将会异步的执行，每当订阅的标的tick发生变化更新，callback回调函数就会被调用一次
+# 本地已有的数据不会触发callback
+
+# 定义的回测函数
+## 回调函数中，data是本次触发回调的数据，只有一条
+def f(data):
+    # print(data)
+
+    code_list = list(data.keys())  # 获取到本次触发的标的代码
+
+    kline_in_callabck = xtdata.get_market_data_ex([], code_list, period=period)  # 在回调中获取klines数据
+    print(kline_in_callabck)
+
+
+for i in code_list:
+    xtdata.subscribe_quote(i, period=period, count=-1, callback=f)  # 订阅时设定回调函数
+
+# 使用回调时，必须要同时使用xtdata.run()来阻塞程序，否则程序运行到最后一行就直接结束退出了。
+xtdata.run()
+
+
+