feat: 提交

llmagent/__init__.py

@@ -5,3 +5,26 @@
 # @File    : __init__.py
 # @Project : reActLLMDemo
 # @Desc    :
+from langchain_openai import ChatOpenAI
+from llmagent.llm_config import LLMConfigLoader
+from llmagent.llm_config import base_conf
+from llmtools.tool_impl import tools
+
+# 初始化LLM模型
+llm_conf = LLMConfigLoader.load(item_name=base_conf.model_form)
+llm = ChatOpenAI(
+    model=llm_conf.model, api_key=llm_conf.api_key,
+    base_url=llm_conf.base_url, max_tokens=llm_conf.max_tokens,
+    temperature=llm_conf.temperature,
+    streaming=llm_conf.streaming
+)
+# 绑定工具
+llm_with_tools = llm.bind_tools(tools)
+
+# 提示词模版
+PROMPT_TEMPLATE = {
+    'SMART_ASSISTANT': {
+        'description': '智能助手',
+        'template': """"""
+    }
+}

llmagent/assistant_graph.py

@@ -0,0 +1,62 @@
+#!/usr/bin/env python
+# -*- coding: utf-8 -*-
+# @Time    : 2025/4/17 15:28
+# @Author  : old-tom
+# @File    : llm_agent
+# @Project : reActLLMDemo
+# @Desc    : 用图构建ReAct
+from typing import Annotated
+
+from langgraph.checkpoint.memory import MemorySaver
+
+from llmagent import llm_with_tools
+from llmtools.tool_impl import tool_node
+from langchain_core.messages import AnyMessage, SystemMessage
+from typing_extensions import TypedDict
+from langgraph.graph import StateGraph, START, END
+from langgraph.graph.message import add_messages
+from langgraph.prebuilt import tools_condition
+from langchain_core.runnables import RunnableConfig
+
+# 内存记忆
+memory = MemorySaver()
+
+
+class AgentState(TypedDict):
+    """
+    状态机
+    add_messages 函数会自动合并message到一个list中，例如HumanMessage\AIMessage
+    """
+    messages: Annotated[list[AnyMessage], add_messages]
+
+
+graph_builder = StateGraph(AgentState)
+
+
+def chat(state: AgentState, config: RunnableConfig):
+    """
+    调用LLM
+    :param state: 状态机
+    :param config: 配置
+    LLM需要从状态机获取message
+    :return:
+    """
+    # 设置系统提示词
+    system_prompt = SystemMessage(
+        "You are a helpful AI assistant, please respond to the users query to the best of your ability!"
+    )
+    return {"messages": [llm_with_tools.invoke([system_prompt] + state["messages"], config)]}
+
+
+# 以下步骤可以替换为预构建的create_react_agent函数
+# LLM节点
+graph_builder.add_node("chat_llm", chat)
+# 工具节点
+graph_builder.add_node("tools", tool_node)
+graph_builder.add_edge(START, "chat_llm")
+graph_builder.add_edge("chat_llm", END)
+# 添加条件边，tools_condition 是官方实现的函数，用于判断是否应该调用tool或者直接结束
+graph_builder.add_conditional_edges("chat_llm", tools_condition)
+graph_builder.add_edge("tools", "chat_llm")
+# checkpointer 是检查点设置
+graph = graph_builder.compile(name='smart_assistant', checkpointer=memory)

llmagent/llm_agent.py

@@ -1,91 +0,0 @@
-#!/usr/bin/env python
-# -*- coding: utf-8 -*-
-# @Time    : 2025/4/17 15:28
-# @Author  : old-tom
-# @File    : llm_agent
-# @Project : reActLLMDemo
-# @Desc    : 代理
-from typing import Annotated
-
-from langgraph.checkpoint.memory import MemorySaver
-from langchain_openai import ChatOpenAI
-from llmagent.llm_config import LLMConfigLoader
-from llmagent.llm_config import base_conf
-from llmtools.tool_impl import tools, tool_node
-from langchain_core.messages import AnyMessage
-from typing_extensions import TypedDict
-from langgraph.graph import StateGraph, START, END
-from langgraph.graph.message import add_messages
-from langgraph.prebuilt import tools_condition
-
-# 内存记忆
-memory = MemorySaver()
-
-# 初始化LLM模型
-llm_conf = LLMConfigLoader.load(item_name=base_conf.model_form)
-llm = ChatOpenAI(
-    model=llm_conf.model, api_key=llm_conf.api_key,
-    base_url=llm_conf.base_url, max_tokens=llm_conf.max_tokens,
-    temperature=llm_conf.temperature,
-    streaming=llm_conf.streaming
-)
-# 绑定工具
-llm_with_tools = llm.bind_tools(tools)
-
-
-class AgentState(TypedDict):
-    """
-    状态机
-    add_messages 函数会自动合并message到一个list中，例如HumanMessage\AIMessage
-    """
-    messages: Annotated[list[AnyMessage], add_messages]
-
-
-graph_builder = StateGraph(AgentState)
-
-
-def chat(state: AgentState):
-    """
-    LLM单轮对话
-    :param state: 状态机
-    LLM需要从状态机获取message
-    :return:
-    """
-    return {"messages": [llm_with_tools.invoke(state["messages"])]}
-
-
-# LLM节点
-graph_builder.add_node("chat_llm", chat)
-# 工具节点
-graph_builder.add_node("tools", tool_node)
-graph_builder.add_edge(START, "chat_llm")
-graph_builder.add_edge("chat_llm", END)
-# 添加条件边，tools_condition 是官方实现的函数，用于判断是否应该调用tool或者直接结束
-graph_builder.add_conditional_edges("chat_llm", tools_condition)
-graph_builder.add_edge("tools", "chat_llm")
-# checkpointer 是检查点设置
-graph = graph_builder.compile(name='语音助手', checkpointer=memory)
-
-
-def stream_graph_updates(user_input: str):
-    config = {"configurable": {"thread_id": "1"}}
-    for chunk, metadata in graph.stream({"messages": [{"role": "user", "content": user_input}]}, config,
-                                        stream_mode='messages'):
-        if chunk.content:
-            print(chunk.content, end='', flush=True)
-    print('\n')
-
-
-while True:
-    try:
-        user_input = input("User: ")
-        if user_input.lower() in ["quit", "exit", "q"]:
-            print("Goodbye!")
-            break
-
-        stream_graph_updates(user_input)
-    except:
-        user_input = "What do you know about LangGraph?"
-        print("User: " + user_input)
-        stream_graph_updates(user_input)
-        break

local_test.py

@@ -0,0 +1,39 @@
+#!/usr/bin/env python
+# -*- coding: utf-8 -*-
+# @Time    : 2025/4/19 18:27
+# @Author  : old-tom
+# @File    : local_test
+# @Project : reActLLMDemo
+# @Desc    : 本地终端运行对话
+import langgraph.store.memory
+from llmagent.assistant_graph import graph
+
+
+def stream_graph_updates(user_input: str):
+    """
+    流式输出
+    :param user_input:
+    :return:
+    """
+    config = {"configurable": {"thread_id": "1"}}
+    for chunk, metadata in graph.stream({"messages": [{"role": "user", "content": user_input}]}, config,
+                                        stream_mode='messages'):
+        if chunk.content:
+            print(chunk.content, end='', flush=True)
+    print('\n')
+
+
+if __name__ == '__main__':
+    while True:
+        try:
+            user_input = input("User: ")
+            if user_input.lower() in ["quit", "exit", "q"]:
+                print("Goodbye!")
+                break
+
+            stream_graph_updates(user_input)
+        except:
+            user_input = "What do you know about LangGraph?"
+            print("User: " + user_input)
+            stream_graph_updates(user_input)
+            break

main.py

@@ -1,6 +0,0 @@
-def main():
-    print("Hello from reactllmdemo!")
-
-
-if __name__ == "__main__":
-    main()

server.py

@@ -0,0 +1,14 @@
+#!/usr/bin/env python
+# -*- coding: utf-8 -*-
+# @Time    : 2025/4/19 18:27
+# @Author  : old-tom
+# @File    : local_test
+# @Project : reActLLMDemo
+# @Desc    : 作为服务端运行
+
+def main():
+    print("Hello from reactllmdemo!")
+
+
+if __name__ == "__main__":
+    main()

vector_db.py

@@ -8,13 +8,13 @@
 import marqo
 
 # 索引名称
-INDEX_NAME = 'test_index'
+INDEX_NAME = 'camera_collection'
 # 初始化marqo
 mq = marqo.Client(url='http://171.92.0.3:8882')
 
 # mq.delete_index(INDEX_NAME)
 
-
+#
 # settings = {
 #     "treatUrlsAndPointersAsImages": False,
 #     "model": "hf/bge-large-zh-v1.5",
@@ -28,6 +28,7 @@ def create_and_set_index():
     :return:
     """
     mq.create_index(INDEX_NAME, model='hf/e5-base-v2')
+    # mq.create_index(INDEX_NAME, settings_dict=settings)
     # 添加文档(测试用)
     mq.index(INDEX_NAME).add_documents([
         {
@@ -625,7 +626,7 @@ def query_vector_db(query):
 
 if __name__ == '__main__':
     # create_and_set_index()
-    rt = query_vector_db('利丰高点')
+    rt = query_vector_db('南卡口')
     # TODO 根据 _score字段 取出相似度最高的结果
     if rt:
         for ele in rt['hits']: