Update agent.py

agent.py

@@ -1,95 +1,213 @@
-from typing import Annotated, Sequence, TypedDict
-from langchain_community.llms import HuggingFaceHub,HuggingFaceEndpoint  
-from langchain_core.messages import BaseMessage, HumanMessage, AIMessage
-from langgraph.graph import StateGraph, END
-from langchain_core.agents import AgentAction, AgentFinish
-from langchain.agents import create_react_agent
-from langchain import hub
-from ai_tools import get_tools  # 导入自定义工具集
-import operator 
-
-class AgentState(TypedDict):
-    messages: Annotated[Sequence[BaseMessage], operator.add]
-    intermediate_steps: Annotated[list, operator.add]
-
-def build_graph():
-    # 1. 初始化模型 - 使用HuggingFace免费接口
-    llm = HuggingFaceEndpoint(
-        endpoint_url="https://api-inference.huggingface.co/models/mistralai/Mistral-7B-Instruct-v0.2",
-        max_new_tokens=500,          # 直接指定
-        temperature=0.1,             # 直接指定
-        repetition_penalty=1.2,      # 直接指定
-        top_p=0.9,                   # 可选参数
-        # 其他参数也可以直接在这里指定
-    )
+"""LangGraph Agent"""
+import os
+from dotenv import load_dotenv
+from langgraph.graph import START, StateGraph, MessagesState
+from langgraph.prebuilt import tools_condition
+from langgraph.prebuilt import ToolNode
+from langchain_google_genai import ChatGoogleGenerativeAI
+from langchain_groq import ChatGroq
+from langchain_huggingface import ChatHuggingFace, HuggingFaceEndpoint, HuggingFaceEmbeddings
+from langchain_community.tools.tavily_search import TavilySearchResults
+from langchain_community.document_loaders import WikipediaLoader
+from langchain_community.document_loaders import ArxivLoader
+from langchain_community.vectorstores import SupabaseVectorStore
+from langchain_core.messages import SystemMessage, HumanMessage
+from langchain_core.tools import tool
+from langchain.tools.retriever import create_retriever_tool
+from supabase.client import Client, create_client
+
+load_dotenv()
+
+@tool
+def multiply(a: int, b: int) -> int:
+    """Multiply two numbers.
+    Args:
+        a: first int
+        b: second int
+    """
+    return a * b
+
+@tool
+def add(a: int, b: int) -> int:
+    """Add two numbers.
     
-    # 2. 创建ReAct代理
-    prompt = hub.pull("hwchase17/react")
-    tools = get_tools()
-    agent = create_react_agent(llm, tools, prompt)
+    Args:
+        a: first int
+        b: second int
+    """
+    return a + b
+
+@tool
+def subtract(a: int, b: int) -> int:
+    """Subtract two numbers.
     
-    # 3. 定义节点行为
-    def agent_node(state: AgentState):
-        input = state["messages"][-1].content
-        result = agent.invoke({
-            "input": input,
-            "intermediate_steps": state["intermediate_steps"]
-        })
-        return {"intermediate_steps": [result]}
+    Args:
+        a: first int
+        b: second int
+    """
+    return a - b
+
+@tool
+def divide(a: int, b: int) -> int:
+    """Divide two numbers.
     
-    def tool_node(state: AgentState):
-        last_step = state["intermediate_steps"][-1]
-        action = last_step[0] if isinstance(last_step, list) else last_step
-        
-        if not isinstance(action, AgentAction):
-            return {"messages": [AIMessage(content="Invalid action format")]}
-        
-        # 执行工具调用
-        tool = next((t for t in tools if t.name == action.tool), None)
-        if not tool:
-            return {"messages": [AIMessage(content=f"Tool {action.tool} not found")]}
-        
-        observation = tool.run(action.tool_input)
-        return {"messages": [AIMessage(content=observation)]}
+    Args:
+        a: first int
+        b: second int
+    """
+    if b == 0:
+        raise ValueError("Cannot divide by zero.")
+    return a / b
+
+@tool
+def modulus(a: int, b: int) -> int:
+    """Get the modulus of two numbers.
     
-    # 4. 构建状态图
-    workflow = StateGraph(AgentState)
-    workflow.add_node("agent", agent_node)
-    workflow.add_node("tool", tool_node)
+    Args:
+        a: first int
+        b: second int
+    """
+    return a % b
+
+@tool
+def wiki_search(query: str) -> str:
+    """Search Wikipedia for a query and return maximum 2 results.
     
-    # 5. 定义边和条件
-    def route_action(state: AgentState):
-        last_step = state["intermediate_steps"][-1]
-        action = last_step[0] if isinstance(last_step, list) else last_step
-        
-        if isinstance(action, AgentFinish):
-            return END
-        return "tool"
+    Args:
+        query: The search query."""
+    search_docs = WikipediaLoader(query=query, load_max_docs=2).load()
+    formatted_search_docs = "\n\n---\n\n".join(
+        [
+            f'<Document source="{doc.metadata["source"]}" page="{doc.metadata.get("page", "")}"/>\n{doc.page_content}\n</Document>'
+            for doc in search_docs
+        ])
+    return {"wiki_results": formatted_search_docs}
+
+@tool
+def web_search(query: str) -> str:
+    """Search Tavily for a query and return maximum 3 results.
     
-    workflow.set_entry_point("agent")
-    workflow.add_conditional_edges(
-        "agent",
-        route_action,
-        {"tool": "tool", END: END}
-    )
-    workflow.add_edge("tool", "agent")
+    Args:
+        query: The search query."""
+    search_docs = TavilySearchResults(max_results=3).invoke(query=query)
+    formatted_search_docs = "\n\n---\n\n".join(
+        [
+            f'<Document source="{doc.metadata["source"]}" page="{doc.metadata.get("page", "")}"/>\n{doc.page_content}\n</Document>'
+            for doc in search_docs
+        ])
+    return {"web_results": formatted_search_docs}
+
+@tool
+def arvix_search(query: str) -> str:
+    """Search Arxiv for a query and return maximum 3 result.
     
-    return workflow.compile()
+    Args:
+        query: The search query."""
+    search_docs = ArxivLoader(query=query, load_max_docs=3).load()
+    formatted_search_docs = "\n\n---\n\n".join(
+        [
+            f'<Document source="{doc.metadata["source"]}" page="{doc.metadata.get("page", "")}"/>\n{doc.page_content[:1000]}\n</Document>'
+            for doc in search_docs
+        ])
+    return {"arvix_results": formatted_search_docs}
+
 
-class BasicAgent:
-    """LangGraph智能体封装"""
-    def __init__(self):
-        print("BasicAgent initialized.")
-        self.graph = build_graph()
+
+# load the system prompt from the file
+with open("system_prompt.txt", "r", encoding="utf-8") as f:
+    system_prompt = f.read()
+
+# System message
+sys_msg = SystemMessage(content=system_prompt)
+
+# build a retriever
+embeddings = HuggingFaceEmbeddings(model_name="sentence-transformers/all-mpnet-base-v2") #  dim=768
+supabase: Client = create_client(
+    os.environ.get("SUPABASE_URL"), 
+    os.environ.get("SUPABASE_SERVICE_KEY"))
+vector_store = SupabaseVectorStore(
+    client=supabase,
+    embedding= embeddings,
+    table_name="documents",
+    query_name="match_documents_langchain",
+)
+create_retriever_tool = create_retriever_tool(
+    retriever=vector_store.as_retriever(),
+    name="Question Search",
+    description="A tool to retrieve similar questions from a vector store.",
+)
+
+
+
+tools = [
+    multiply,
+    add,
+    subtract,
+    divide,
+    modulus,
+    wiki_search,
+    web_search,
+    arvix_search,
+]
+
+# Build graph function
+def build_graph(provider: str = "groq"):
+    """Build the graph"""
+    # Load environment variables from .env file
+    if provider == "google":
+        # Google Gemini
+        llm = ChatGoogleGenerativeAI(model="gemini-2.0-flash", temperature=0)
+    elif provider == "groq":
+        # Groq https://console.groq.com/docs/models
+        llm = ChatGroq(model="qwen-qwq-32b", temperature=0) # optional : qwen-qwq-32b gemma2-9b-it
+    elif provider == "huggingface":
+        # TODO: Add huggingface endpoint
+        llm = ChatHuggingFace(
+            llm=HuggingFaceEndpoint(
+                url="https://api-inference.huggingface.co/models/Meta-DeepLearning/llama-2-7b-chat-hf",
+                temperature=0,
+            ),
+        )
+    else:
+        raise ValueError("Invalid provider. Choose 'google', 'groq' or 'huggingface'.")
+    # Bind tools to LLM
+    llm_with_tools = llm.bind_tools(tools)
+
+    # Node
+    def assistant(state: MessagesState):
+        """Assistant node"""
+        return {"messages": [llm_with_tools.invoke(state["messages"])]}
     
-    def __call__(self, question: str) -> str:
-        print(f"Agent received question: {question[:50]}...")
-        messages = [HumanMessage(content=question)]
-        result = self.graph.invoke({
-            "messages": messages,
-            "intermediate_steps": []
-        })
-        
-        # 提取最终答案
-        final_message = result["messages"][-1].content
-        return final_message.strip()
+    def retriever(state: MessagesState):
+        """Retriever node"""
+        similar_question = vector_store.similarity_search(state["messages"][0].content)
+        example_msg = HumanMessage(
+            content=f"Here I provide a similar question and answer for reference: \n\n{similar_question[0].page_content}",
+        )
+        return {"messages": [sys_msg] + state["messages"] + [example_msg]}
+
+    builder = StateGraph(MessagesState)
+    builder.add_node("retriever", retriever)
+    builder.add_node("assistant", assistant)
+    builder.add_node("tools", ToolNode(tools))
+    builder.add_edge(START, "retriever")
+    builder.add_edge("retriever", "assistant")
+    builder.add_conditional_edges(
+        "assistant",
+        tools_condition,
+    )
+    builder.add_edge("tools", "assistant")
+
+    # Compile graph
+    return builder.compile()
+
+# test
+if __name__ == "__main__":
+    question = "When was a picture of St. Thomas Aquinas first added to the Wikipedia page on the Principle of double effect?"
+    # Build the graph
+    graph = build_graph(provider="groq")
+    # Run the graph
+    messages = [HumanMessage(content=question)]
+    messages = graph.invoke({"messages": messages})
+    for m in messages["messages"]:
+        m.pretty_print()