Update app.py

app.py

@@ -2,25 +2,24 @@ import os
 import gradio as gr
 import requests
 import pandas as pd
-import time
 
-# --- New Imports for Tool-Calling Agent ---
+# --- Updated Imports for Modern LangChain (v1.0+) ---
 from langchain_google_genai import ChatGoogleGenerativeAI
 from langchain_community.tools import DuckDuckGoSearchRun
-from langchain.agents import AgentExecutor, create_tool_calling_agent
-from langchain_core.prompts import ChatPromptTemplate
+from langgraph.prebuilt import create_react_agent
+from langchain_core.messages import SystemMessage
 
 # --- Constants ---
 DEFAULT_API_URL = "https://agents-course-unit4-scoring.hf.space"
 
-# --- Refactored Gemini Agent Definition ---
+# --- Modern LangGraph Gemini Agent ---
 class GeminiAgent:
     def __init__(self):
         api_key = os.getenv("GEMINI_API_KEY")
         if not api_key:
             raise ValueError("GEMINI_API_KEY not set")
 
-        # Initialize Gemini 2.5 Flash for fast, accurate tool calling
+        # Initialize Gemini 2.5 Flash
         self.llm = ChatGoogleGenerativeAI(
             model="gemini-2.5-flash",
             temperature=0,
@@ -31,38 +30,33 @@ class GeminiAgent:
         self.search_tool = DuckDuckGoSearchRun()
         self.tools = [self.search_tool]
 
-        # Define the Agentic Prompt
-        prompt = ChatPromptTemplate.from_messages([
-            ("system", """You are an expert assistant for the GAIA benchmark. 
-            You must use your tools to find accurate, up-to-date information before answering.
-            Do not guess. If you need to perform math, search for the formula or calculation.
-            Provide ONLY a short, factual answer (e.g., a specific number, name, or exact phrase). 
-            No explanations, just the direct answer."""),
-            ("human", "{input}"),
-            ("placeholder", "{agent_scratchpad}"),
-        ])
+        # Define the System Prompt
+        system_prompt = """You are an expert assistant for the GAIA benchmark. 
+        You must use your tools to find accurate, up-to-date information before answering.
+        Do not guess. If you need to perform math, search for the formula or calculation.
+        Provide ONLY a short, factual answer (e.g., a specific number, name, or exact phrase). 
+        No explanations, just the direct answer."""
 
-        # Create the Tool Calling Agent and Executor
-        self.agent = create_tool_calling_agent(self.llm, self.tools, prompt)
-        self.agent_executor = AgentExecutor(
-            agent=self.agent, 
-            tools=self.tools, 
-            verbose=True, # Set to False to reduce logs
-            max_iterations=5, # Prevent infinite loops
-            handle_parsing_errors=True
+        # Create the LangGraph Agent (replaces AgentExecutor)
+        self.agent_executor = create_react_agent(
+            self.llm, 
+            self.tools,
+            state_modifier=SystemMessage(content=system_prompt)
         )
 
-        print("Gemini Tool-Calling Agent initialized with Gemini 2.5 Flash")
+        print("LangGraph Agent initialized with Gemini 2.5 Flash")
 
     def __call__(self, question: str) -> str:
         print(f"Agent processing question: {question[:50]}...")
         
         try:
-            # We no longer need time.sleep(6) because Gemini 2.5 Flash handles rate limits better,
-            # and the agent executor handles the pacing of tool calls natively.
-            response = self.agent_executor.invoke({"input": question})
+            # LangGraph expects input formatted as a list of messages
+            response = self.agent_executor.invoke({
+                "messages": [("user", question)]
+            })
             
-            answer = response.get("output", "").strip()
+            # The final answer is the content of the last message in the sequence
+            answer = response["messages"][-1].content.strip()
             
             if not answer:
                 answer = "0"