Initial commit: 国土空间规划课程智能体 v1.0

单容器 Docker 架构的国土空间规划课程智能问答系统，集成 FastAPI 后端与 Next.js 前端。 Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
2026-05-22 09:40:18 +08:00
commit ddbb79b9f6
167 changed files with 44147 additions and 0 deletions
@@ -0,0 +1,129 @@
+"""
+LangChain 1.0 对话链（Normal Mode）
+"""
+from typing import List, Dict, Any, Optional
+from langchain_core.runnables import RunnablePassthrough
+from langchain_core.prompts import ChatPromptTemplate, MessagesPlaceholder
+from langchain_core.messages import HumanMessage, AIMessage, SystemMessage
+from langchain_core.output_parsers import StrOutputParser
+
+from ..llm.siliconflow import get_llm_client
+
+
+class ConversationChain:
+    """标准对话链（LangChain 1.0）"""
+    
+    def __init__(
+        self,
+        system_prompt: Optional[str] = None,
+        model: Optional[str] = None
+    ):
+        """初始化对话链
+        
+        Args:
+            system_prompt: 系统提示词
+            model: 可选的模型名称，如 deepseek-ai/DeepSeek-V3, Qwen/QwQ-32B
+        """
+        self.llm = get_llm_client(model=model).llm
+        self.system_prompt = system_prompt or "你是一个专业的国土空间规划知识问答助手。请基于你的知识回答用户的问题。"
+        
+        # 创建带历史的Prompt模板
+        self.prompt = ChatPromptTemplate.from_messages([
+            SystemMessage(content=self.system_prompt),
+            MessagesPlaceholder(variable_name="chat_history"),
+            ("human", "{question}")
+        ])
+        
+        # 创建对话链
+        self.chain = self._create_chain()
+    
+    def _create_chain(self):
+        """创建对话链"""
+        return (
+            self.prompt
+            | self.llm
+            | StrOutputParser()
+        )
+    
+    def invoke(self, question: str, chat_history: List[Dict] = None) -> Dict[str, Any]:
+        """同步调用"""
+        history_messages = self._format_history(chat_history or [])
+        answer = self.chain.invoke({
+            "question": question,
+            "chat_history": history_messages
+        })
+        
+        return {
+            "answer": answer,
+            "sources": [],
+            "metadata": {
+                "mode": "normal",
+                "has_history": bool(chat_history)
+            }
+        }
+    
+    async def ainvoke(self, question: str, chat_history: List[Dict] = None) -> Dict[str, Any]:
+        """异步调用"""
+        history_messages = self._format_history(chat_history or [])
+        answer = await self.chain.ainvoke({
+            "question": question,
+            "chat_history": history_messages
+        })
+        
+        return {
+            "answer": answer,
+            "sources": [],
+            "metadata": {
+                "mode": "normal",
+                "has_history": bool(chat_history)
+            }
+        }
+    
+    async def astream_with_thinking(self, question: str, chat_history: List[Dict] = None):
+        """流式调用（包含思考过程）"""
+        import time
+        
+        # 思考阶段
+        start_time = time.time()
+        yield {"type": "thinking", "stage": "understanding", "message": "正在理解问题..."}
+        
+        # 准备历史
+        history_messages = self._format_history(chat_history or [])
+        
+        yield {"type": "thinking", "stage": "generating", "message": "正在生成回答..."}
+        
+        # 流式生成
+        async for chunk in self.chain.astream({
+            "question": question,
+            "chat_history": history_messages
+        }):
+            yield {"type": "chunk", "content": chunk}
+        
+        # 完成
+        total_time = time.time() - start_time
+        yield {
+            "type": "complete",
+            "metadata": {
+                "total_time": round(total_time, 2)
+            }
+        }
+    
+    def _format_history(self, chat_history: List[Dict]) -> List:
+        """格式化聊天历史为LangChain消息格式"""
+        messages = []
+        for msg in chat_history:
+            if msg["role"] == "user":
+                messages.append(HumanMessage(content=msg["content"]))
+            elif msg["role"] == "assistant":
+                messages.append(AIMessage(content=msg["content"]))
+        return messages
+
+
+def create_conversation_chain(system_prompt: Optional[str] = None, model: Optional[str] = None) -> ConversationChain:
+    """创建对话链实例
+    
+    Args:
+        system_prompt: 系统提示词
+        model: 可选的模型名称，如 deepseek-ai/DeepSeek-V3, Qwen/QwQ-32B
+    """
+    return ConversationChain(system_prompt=system_prompt, model=model)