Update agent.py

agent.py

@@ -0,0 +1,63 @@
+import os
+from dotenv import load_dotenv
+from langgraph.graph import START, END, StateGraph, MessagesState
+from langgraph.prebuilt import ToolNode, tools_condition
+from langchain_huggingface import ChatHuggingFace, HuggingFaceEndpoint
+from langchain_core.tools import tool
+
+load_dotenv()
+
+# --- 1. Define Basic Tools ---
+@tool
+def add(a: int, b: int) -> int:
+    """Add two numbers."""
+    return a + b
+
+@tool
+def multiply(a: int, b: int) -> int:
+    """Multiply two numbers."""
+    return a * b
+
+tools = [add, multiply]
+
+# --- 2. Initialize the Hugging Face LLM ---
+def get_llm():
+    # We use HuggingFaceEndpoint to connect to the model API
+    # You can change the repo_id to other models like "google/gemma-2-9b-it"
+    llm_engine = HuggingFaceEndpoint(
+        repo_id="meta-llama/Llama-3.3-70B-Instruct", 
+        task="text-generation",
+        huggingfacehub_api_token=os.getenv("HUGGINGFACEHUB_API_TOKEN")
+    )
+    
+    # Wrap it in ChatHuggingFace to make it compatible with LangGraph tools
+    model = ChatHuggingFace(llm=llm_engine)
+    return model.bind_tools(tools)
+
+# --- 3. Define the Logic Nodes ---
+def assistant(state: MessagesState):
+    llm = get_llm()
+    # The agent looks at the message history and decides the next move
+    return {"messages": [llm.invoke(state["messages"])]}
+
+# --- 4. Build the Graph ---
+def build_graph():
+    builder = StateGraph(MessagesState)
+
+    # Add Nodes
+    builder.add_node("assistant", assistant)
+    builder.add_node("tools", ToolNode(tools))
+
+    # Define Flow
+    builder.add_edge(START, "assistant")
+    
+    # The conditional edge checks if the LLM called a tool
+    builder.add_conditional_edges(
+        "assistant",
+        tools_condition, # If tool called -> "tools", else -> END
+    )
+    
+    # After using a tool, always go back to the assistant to summarize
+    builder.add_edge("tools", "assistant")
+
+    return builder.compile()