TL

backend/app/api/endpoints/ai_analyze.py

@@ -10,6 +10,7 @@ import os
 
 from app.services.excel_ai_service import excel_ai_service
 from app.services.markdown_ai_service import markdown_ai_service
+from app.services.template_fill_service import template_fill_service
 
 logger = logging.getLogger(__name__)
 
@@ -340,3 +341,74 @@ async def get_markdown_outline(
     except Exception as e:
         logger.error(f"获取 Markdown 大纲失败: {str(e)}")
         raise HTTPException(status_code=500, detail=f"获取大纲失败: {str(e)}")
+
+
+@router.post("/analyze/txt")
+async def analyze_txt(
+    file: UploadFile = File(...),
+):
+    """
+    上传并使用 AI 分析 TXT 文本文件，提取结构化数据
+
+    将非结构化文本转换为结构化表格数据，便于后续填表使用
+
+    Args:
+        file: 上传的 TXT 文件
+
+    Returns:
+        dict: 分析结果，包含结构化表格数据
+    """
+    if not file.filename:
+        raise HTTPException(status_code=400, detail="文件名为空")
+
+    file_ext = file.filename.split('.')[-1].lower()
+    if file_ext not in ['txt', 'text']:
+        raise HTTPException(
+            status_code=400,
+            detail=f"不支持的文件类型: {file_ext}，仅支持 .txt"
+        )
+
+    try:
+        # 读取文件内容
+        content = await file.read()
+
+        # 保存到临时文件
+        with tempfile.NamedTemporaryFile(mode='wb', suffix='.txt', delete=False) as tmp:
+            tmp.write(content)
+            tmp_path = tmp.name
+
+        try:
+            logger.info(f"开始 AI 分析 TXT 文件: {file.filename}")
+
+            # 使用 template_fill_service 的 AI 分析方法
+            result = await template_fill_service.analyze_txt_with_ai(
+                content=content.decode('utf-8', errors='replace'),
+                filename=file.filename
+            )
+
+            if result:
+                logger.info(f"TXT AI 分析成功: {file.filename}")
+                return {
+                    "success": True,
+                    "filename": file.filename,
+                    "structured_data": result
+                }
+            else:
+                logger.warning(f"TXT AI 分析返回空结果: {file.filename}")
+                return {
+                    "success": False,
+                    "filename": file.filename,
+                    "error": "AI 分析未能提取到结构化数据",
+                    "structured_data": None
+                }
+
+        finally:
+            # 清理临时文件
+            if os.path.exists(tmp_path):
+                os.unlink(tmp_path)
+
+    except HTTPException:
+        raise
+    except Exception as e:
+        logger.error(f"TXT AI 分析过程中出错: {str(e)}")
+        raise HTTPException(status_code=500, detail=f"分析失败: {str(e)}")

backend/app/main.py

@@ -1,6 +1,13 @@
 """
 FastAPI 应用主入口
 """
+# ========== 压制 MongoDB 疯狂刷屏日志 ==========
+import logging
+logging.getLogger("pymongo").setLevel(logging.WARNING)
+logging.getLogger("pymongo.topology").setLevel(logging.WARNING)
+logging.getLogger("urllib3").setLevel(logging.WARNING)
+# ==============================================
+
 import logging
 import logging.handlers
 import sys

backend/app/services/llm_service.py

@@ -42,33 +42,86 @@ class LLMService:
             "Content-Type": "application/json"
         }
 
+        # DeepSeek API temperature 范围: (0, 2]
+        if temperature < 0.01:
+            temperature = 0.01
+        elif temperature > 2.0:
+            temperature = 2.0
+
         payload = {
             "model": self.model_name,
             "messages": messages,
             "temperature": temperature
         }
 
+        # DeepSeek API 限制 max_tokens 范围
         if max_tokens:
+            if max_tokens > 8192:
+                max_tokens = 8192
             payload["max_tokens"] = max_tokens
 
+        # 移除不兼容的参数
+        for key in ["stream", "stop", "presence_penalty", "frequency_penalty", "logit_bias"]:
+            kwargs.pop(key, None)
+
         # 添加其他参数
         payload.update(kwargs)
 
+        # 验证消息格式
+        validated_messages = []
+        for i, msg in enumerate(messages):
+            role = msg.get("role", "")
+            content = msg.get("content", "")
+
+            # 确保 content 是字符串
+            if not isinstance(content, str):
+                logger.warning(f"消息[{i}] content 不是字符串类型: {type(content)}，转换为字符串")
+                content = str(content)
+
+            # 确保 role 有效
+            if role not in ["system", "user", "assistant"]:
+                logger.warning(f"消息[{i}] role 无效: {role}，跳过")
+                continue
+
+            validated_messages.append({"role": role, "content": content})
+
+        payload["messages"] = validated_messages
+        logger.info(f"验证后消息数量: {len(validated_messages)}")
+
         try:
-            async with httpx.AsyncClient(timeout=60.0) as client:
+            logger.info(f"LLM API 请求: model={self.model_name}, base_url={self.base_url}, temperature={temperature}, max_tokens={max_tokens}")
+            logger.info(f"消息数量: {len(messages)}")
+            total_content_len = sum(len(msg.get('content', '')) for msg in messages)
+            logger.info(f"总内容长度: {total_content_len}")
+
+            async with httpx.AsyncClient(timeout=120.0) as client:
                 response = await client.post(
                     f"{self.base_url}/chat/completions",
                     headers=headers,
                     json=payload
                 )
+
+                logger.info(f"LLM API 响应状态: {response.status_code}")
+
+                if response.status_code != 200:
+                    error_text = response.text
+                    logger.error(f"LLM API 错误响应: {error_text}")
+                    # 尝试解析错误详情
+                    try:
+                        error_json = response.json()
+                        error_msg = error_json.get("error", {}).get("message", error_text)
+                        logger.error(f"错误详情: {error_msg}")
+                    except:
+                        pass
+
                 response.raise_for_status()
                 return response.json()
 
         except httpx.HTTPStatusError as e:
-            logger.error(f"LLM API 请求失败: {e.response.status_code} - {e.response.text}")
+            logger.error(f"LLM API HTTP 错误: {e.response.status_code} - {e.response.text}")
             raise
         except Exception as e:
-            logger.error(f"LLM API 调用异常: {str(e)}")
+            logger.error(f"LLM API 调用异常: {str(e)}", exc_info=True)
             raise
 
     def extract_message_content(self, response: Dict[str, Any]) -> str:
@@ -111,6 +164,10 @@ class LLMService:
             "Content-Type": "application/json"
         }
 
+        # DeepSeek API 限制
+        if max_tokens and max_tokens > 8192:
+            max_tokens = 8192
+
         payload = {
             "model": self.model_name,
             "messages": messages,
@@ -121,9 +178,14 @@ class LLMService:
         if max_tokens:
             payload["max_tokens"] = max_tokens
 
+        # 移除不兼容的参数
+        for key in ["stop", "presence_penalty", "frequency_penalty", "logit_bias"]:
+            kwargs.pop(key, None)
         payload.update(kwargs)
 
         try:
+            logger.info(f"LLM 流式 API 请求: model={self.model_name}, max_tokens={max_tokens}")
+
             async with httpx.AsyncClient(timeout=120.0) as client:
                 async with client.stream(
                     "POST",
@@ -131,9 +193,14 @@ class LLMService:
                     headers=headers,
                     json=payload
                 ) as response:
+                    if response.status_code != 200:
+                        error_text = await response.aread()
+                        logger.error(f"LLM 流式 API 错误: {response.status_code} - {error_text}")
+                        response.raise_for_status()
+
                     async for line in response.aiter_lines():
                         if line.startswith("data: "):
-                            data = line[6:]  # Remove "data: " prefix
+                            data = line[6:]
                             if data == "[DONE]":
                                 break
                             try:
@@ -149,7 +216,7 @@ class LLMService:
             logger.error(f"LLM 流式 API 请求失败: {e.response.status_code}")
             raise
         except Exception as e:
-            logger.error(f"LLM 流式 API 调用异常: {str(e)}")
+            logger.error(f"LLM 流式 API 调用异常: {str(e)}", exc_info=True)
             raise
 
     async def analyze_excel_data(

frontend/src/db/backend-api.ts

@@ -1204,6 +1204,48 @@ export const aiApi = {
     }
   },
 
+  /**
+   * 上传并使用 AI 分析 TXT 文本文件，提取结构化数据
+   */
+  async analyzeTxt(
+    file: File
+  ): Promise<{
+    success: boolean;
+    filename?: string;
+    structured_data?: {
+      table?: {
+        columns?: string[];
+        rows?: string[][];
+      };
+      summary?: string;
+      key_value_pairs?: Array<{ key: string; value: string }>;
+      numeric_data?: Array<{ name: string; value: number; unit?: string }>;
+    };
+    error?: string;
+  }> {
+    const formData = new FormData();
+    formData.append('file', file);
+
+    const url = `${BACKEND_BASE_URL}/ai/analyze/txt`;
+
+    try {
+      const response = await fetch(url, {
+        method: 'POST',
+        body: formData,
+      });
+
+      if (!response.ok) {
+        const error = await response.json();
+        throw new Error(error.detail || 'TXT AI 分析失败');
+      }
+
+      return await response.json();
+    } catch (error) {
+      console.error('TXT AI 分析失败:', error);
+      throw error;
+    }
+  },
+
   /**
    * 生成统计信息和图表
    */