import os import gradio as gr import requests import inspect import pandas as pd # (Keep Constants as is) # --- Constants --- DEFAULT_API_URL = "https://agents-course-unit4-scoring.hf.space" # --- Basic Agent Definition --- # ============================================================================== # 1. IMPORTS AND SETUP # ============================================================================== import os from dotenv import load_dotenv from typing import TypedDict, Annotated, List # LangChain and LangGraph imports from langchain_huggingface import HuggingFaceEndpoint from langchain_community.tools.tavily_search import TavilySearchResults from langchain_experimental.tools import PythonREPLTool from langchain_core.messages import BaseMessage, HumanMessage from langchain_core.prompts import ChatPromptTemplate from langgraph.graph import StateGraph, END from langgraph.prebuilt import ToolNode # ============================================================================== # 2. LOAD API KEYS AND DEFINE TOOLS # ============================================================================== load_dotenv() hf_token = os.getenv("HF_TOKEN") tavily_api_key = os.getenv("TAVILY_API_KEY") if not hf_token or not tavily_api_key: # This will show a clear error in the logs if keys are missing raise ValueError("HF_TOKEN or TAVILY_API_KEY not set. Please add them to your Space secrets.") os.environ["TAVILY_API_KEY"] = tavily_api_key # The agent's tools tools = [TavilySearchResults(max_results=3, description="A search engine for finding up-to-date information on the web."), PythonREPLTool()] tool_node = ToolNode(tools) # ============================================================================== # 3. CONFIGURE THE LLM (THE "BRAIN") # ============================================================================== # The model we'll use as the agent's brain repo_id = "meta-llama/Meta-Llama-3-8B-Instruct" # The system prompt gives the agent its mission and instructions SYSTEM_PROMPT = """You are a highly capable AI agent named 'GAIA-Solver'. Your mission is to accurately answer complex questions. **Your Instructions:** 1. **Analyze:** Carefully read the user's question to understand all parts of what is being asked. 2. **Plan:** Think step-by-step. Break the problem into smaller tasks. Decide which tool is best for each task. (e.g., use 'tavily_search_results_json' for web searches, use 'python_repl' for calculations or code execution). 3. **Execute:** Call ONE tool at a time. 4. **Observe & Reason:** After getting a tool's result, observe it. Decide if you have the final answer or if you need to use another tool. 5. **Final Answer:** Once you are confident, provide a clear, direct, and concise final answer. Do not include your thought process in the final answer. """ # Initialize the LLM endpoint llm = HuggingFaceEndpoint( repo_id=repo_id, huggingfacehub_api_token=hf_token, temperature=0, # Set to 0 for deterministic, less random output max_new_tokens=2048, ) # ============================================================================== # 4. BUILD THE LANGGRAPH AGENT # ============================================================================== # Define the Agent's State (its memory) class AgentState(TypedDict): messages: Annotated[List[BaseMessage], lambda x, y: x + y] # This is a more robust way to combine the prompt, model, and tool binding # It ensures the system prompt is always used. llm_with_tools = llm.bind_tools(tools) # Define the Agent Node def agent_node(state): # Get the last message to pass to the model last_message = state['messages'][-1] # Prepend the system prompt to every call prompt_with_system = [ HumanMessage(content=SYSTEM_PROMPT, name="system_prompt"), last_message ] response = llm_with_tools.invoke(prompt_with_system) return {"messages": [response]} # Define the Edge Logic def should_continue(state): last_message = state["messages"][-1] if last_message.tool_calls: return "tools" # Route to the tool node return END # End the process # Assemble the graph workflow = StateGraph(AgentState) workflow.add_node("agent", agent_node) workflow.add_node("tools", tool_node) workflow.set_entry_point("agent") workflow.add_conditional_edges( "agent", should_continue, {"tools": "tools", "end": END}, ) workflow.add_edge("tools", "agent") # Compile the graph into a runnable app app = workflow.compile() # ============================================================================== # 5. THE BASICAGENT CLASS (FOR THE TEST HARNESS) # This MUST be at the end, after `app` is defined. # ============================================================================== class BasicAgent: """ This is the agent class that the GAIA test harness will use. """ def __init__(self): # The compiled LangGraph app is our agent executor self.agent_executor = app def run(self, question: str) -> str: """ This method is called by the test script with each question. It runs the LangGraph agent and returns the final answer. """ print(f"Agent received question (first 80 chars): {question[:80]}...") try: # Format the input for our graph inputs = {"messages": [HumanMessage(content=question)]} # Stream the response to get the final answer final_response = "" for s in self.agent_executor.stream(inputs, {"recursion_limit": 15}): if "agent" in s: # The final answer is the content of the last message from the agent node if s["agent"]["messages"][-1].content: final_response = s["agent"]["messages"][-1].content # A fallback in case the agent finishes without a clear message if not final_response: final_response = "Agent finished but did not produce a final answer." print(f"Agent returning final answer (first 80 chars): {final_response[:80]}...") return final_response except Exception as e: print(f"An error occurred in agent execution: {e}") return f"Error: {e}" # ----- THIS IS WERE YOU CAN BUILD WHAT YOU WANT ------ # class BasicAgent: # def __init__(self): # print("BasicAgent initialized.") # def __call__(self, question: str) -> str: # print(f"Agent received question (first 50 chars): {question[:50]}...") # fixed_answer = "This is a default answer." # print(f"Agent returning fixed answer: {fixed_answer}") # return fixed_answer def run_and_submit_all( profile: gr.OAuthProfile | None): """ Fetches all questions, runs the BasicAgent on them, submits all answers, and displays the results. """ # --- Determine HF Space Runtime URL and Repo URL --- space_id = os.getenv("SPACE_ID") # Get the SPACE_ID for sending link to the code if profile: username= f"{profile.username}" print(f"User logged in: {username}") else: print("User not logged in.") return "Please Login to Hugging Face with the button.", None api_url = DEFAULT_API_URL questions_url = f"{api_url}/questions" submit_url = f"{api_url}/submit" # 1. Instantiate Agent ( modify this part to create your agent) try: agent = BasicAgent() except Exception as e: print(f"Error instantiating agent: {e}") return f"Error initializing agent: {e}", None # In the case of an app running as a hugging Face space, this link points toward your codebase ( usefull for others so please keep it public) agent_code = f"https://huggingface.co/spaces/{space_id}/tree/main" print(agent_code) # 2. Fetch Questions print(f"Fetching questions from: {questions_url}") try: response = requests.get(questions_url, timeout=15) response.raise_for_status() questions_data = response.json() if not questions_data: print("Fetched questions list is empty.") return "Fetched questions list is empty or invalid format.", None print(f"Fetched {len(questions_data)} questions.") except requests.exceptions.RequestException as e: print(f"Error fetching questions: {e}") return f"Error fetching questions: {e}", None except requests.exceptions.JSONDecodeError as e: print(f"Error decoding JSON response from questions endpoint: {e}") print(f"Response text: {response.text[:500]}") return f"Error decoding server response for questions: {e}", None except Exception as e: print(f"An unexpected error occurred fetching questions: {e}") return f"An unexpected error occurred fetching questions: {e}", None # 3. Run your Agent results_log = [] answers_payload = [] print(f"Running agent on {len(questions_data)} questions...") for item in questions_data: task_id = item.get("task_id") question_text = item.get("question") if not task_id or question_text is None: print(f"Skipping item with missing task_id or question: {item}") continue try: submitted_answer = agent(question_text) answers_payload.append({"task_id": task_id, "submitted_answer": submitted_answer}) results_log.append({"Task ID": task_id, "Question": question_text, "Submitted Answer": submitted_answer}) except Exception as e: print(f"Error running agent on task {task_id}: {e}") results_log.append({"Task ID": task_id, "Question": question_text, "Submitted Answer": f"AGENT ERROR: {e}"}) if not answers_payload: print("Agent did not produce any answers to submit.") return "Agent did not produce any answers to submit.", pd.DataFrame(results_log) # 4. Prepare Submission submission_data = {"username": username.strip(), "agent_code": agent_code, "answers": answers_payload} status_update = f"Agent finished. Submitting {len(answers_payload)} answers for user '{username}'..." print(status_update) # 5. Submit print(f"Submitting {len(answers_payload)} answers to: {submit_url}") try: response = requests.post(submit_url, json=submission_data, timeout=60) response.raise_for_status() result_data = response.json() final_status = ( f"Submission Successful!\n" f"User: {result_data.get('username')}\n" f"Overall Score: {result_data.get('score', 'N/A')}% " f"({result_data.get('correct_count', '?')}/{result_data.get('total_attempted', '?')} correct)\n" f"Message: {result_data.get('message', 'No message received.')}" ) print("Submission successful.") results_df = pd.DataFrame(results_log) return final_status, results_df except requests.exceptions.HTTPError as e: error_detail = f"Server responded with status {e.response.status_code}." try: error_json = e.response.json() error_detail += f" Detail: {error_json.get('detail', e.response.text)}" except requests.exceptions.JSONDecodeError: error_detail += f" Response: {e.response.text[:500]}" status_message = f"Submission Failed: {error_detail}" print(status_message) results_df = pd.DataFrame(results_log) return status_message, results_df except requests.exceptions.Timeout: status_message = "Submission Failed: The request timed out." print(status_message) results_df = pd.DataFrame(results_log) return status_message, results_df except requests.exceptions.RequestException as e: status_message = f"Submission Failed: Network error - {e}" print(status_message) results_df = pd.DataFrame(results_log) return status_message, results_df except Exception as e: status_message = f"An unexpected error occurred during submission: {e}" print(status_message) results_df = pd.DataFrame(results_log) return status_message, results_df # --- Build Gradio Interface using Blocks --- with gr.Blocks() as demo: gr.Markdown("# Basic Agent Evaluation Runner") gr.Markdown( """ **Instructions:** 1. Please clone this space, then modify the code to define your agent's logic, the tools, the necessary packages, etc ... 2. Log in to your Hugging Face account using the button below. This uses your HF username for submission. 3. Click 'Run Evaluation & Submit All Answers' to fetch questions, run your agent, submit answers, and see the score. --- **Disclaimers:** Once clicking on the "submit button, it can take quite some time ( this is the time for the agent to go through all the questions). This space provides a basic setup and is intentionally sub-optimal to encourage you to develop your own, more robust solution. For instance for the delay process of the submit button, a solution could be to cache the answers and submit in a seperate action or even to answer the questions in async. """ ) gr.LoginButton() run_button = gr.Button("Run Evaluation & Submit All Answers") status_output = gr.Textbox(label="Run Status / Submission Result", lines=5, interactive=False) # Removed max_rows=10 from DataFrame constructor results_table = gr.DataFrame(label="Questions and Agent Answers", wrap=True) run_button.click( fn=run_and_submit_all, outputs=[status_output, results_table] ) if __name__ == "__main__": print("\n" + "-"*30 + " App Starting " + "-"*30) # Check for SPACE_HOST and SPACE_ID at startup for information space_host_startup = os.getenv("SPACE_HOST") space_id_startup = os.getenv("SPACE_ID") # Get SPACE_ID at startup if space_host_startup: print(f"✅ SPACE_HOST found: {space_host_startup}") print(f" Runtime URL should be: https://{space_host_startup}.hf.space") else: print("ℹ️ SPACE_HOST environment variable not found (running locally?).") if space_id_startup: # Print repo URLs if SPACE_ID is found print(f"✅ SPACE_ID found: {space_id_startup}") print(f" Repo URL: https://huggingface.co/spaces/{space_id_startup}") print(f" Repo Tree URL: https://huggingface.co/spaces/{space_id_startup}/tree/main") else: print("ℹ️ SPACE_ID environment variable not found (running locally?). Repo URL cannot be determined.") print("-"*(60 + len(" App Starting ")) + "\n") print("Launching Gradio Interface for Basic Agent Evaluation...") demo.launch(debug=True, share=False) ########################################### # import os # import gradio as gr # import requests # import inspect # import pandas as pd # from dotenv import load_dotenv # from typing import TypedDict, Annotated, List # # ============================================================================== # # PART 1: YOUR AGENT'S LOGIC AND DEFINITION # # ============================================================================== # # LangChain and LangGraph imports # from langchain_huggingface import HuggingFaceEndpoint # # NEW: Import TavilySearch from the new package # from langchain_tavily import TavilySearch # from langchain_experimental.tools import PythonREPLTool # from langchain_core.messages import BaseMessage, HumanMessage # from langgraph.graph import StateGraph, END # from langgraph.prebuilt import ToolNode # # NEW: Import the compatible agent constructor and prompt hub # from langchain.agents import create_tool_calling_agent # from langchain import hub # # Load API keys from .env file or Space secrets # load_dotenv() # hf_token = os.getenv("HF_TOKEN") # tavily_api_key = os.getenv("TAVILY_API_KEY") # if tavily_api_key: # os.environ["TAVILY_API_KEY"] = tavily_api_key # else: # print("Warning: TAVILY_API_KEY not found. Web search tool will not work.") # # --- Define Agent Tools --- # # NEW: Using TavilySearch from the correct package # tools = [ # TavilySearch(max_results=3, description="A search engine for finding up-to-date information on the web."), # PythonREPLTool() # ] # tool_node = ToolNode(tools) # # --- Configure the LLM "Brain" --- # repo_id = "meta-llama/Meta-Llama-3-8B-Instruct" # llm = HuggingFaceEndpoint( # repo_id=repo_id, # huggingfacehub_api_token=hf_token, # temperature=0, # max_new_tokens=2048, # ) # # --- THE FIX: Create Agent with a Compatible Method --- # # REMOVED: llm_with_tools = llm.bind_tools(tools) # # This was causing the error. # # NEW: We pull a pre-made prompt that knows how to handle tool calls. # prompt = hub.pull("hwchase17/react-json") # # NEW: We use `create_tool_calling_agent`. This function correctly combines the LLM, # # the tools, and the prompt, without needing the .bind_tools() method. # agent_runnable = create_tool_calling_agent(llm, tools, prompt) # # --- Build the LangGraph Agent --- # class AgentState(TypedDict): # # The 'messages' key is no longer used, 'input' and 'agent_outcome' are standard for this agent type # input: str # chat_history: list[BaseMessage] # agent_outcome: dict # # NEW: The agent_node is much simpler now. It just calls the runnable we created. # def agent_node(state): # outcome = agent_runnable.invoke(state) # return {"agent_outcome": outcome} # def tool_node_executor(state): # # The agent_runnable provides tool calls in a specific format. We execute them. # tool_calls = state["agent_outcome"].tool_calls # tool_outputs = [] # for tool_call in tool_calls: # tool_name = tool_call["name"] # tool_to_call = {tool.name: tool for tool in tools}[tool_name] # tool_output = tool_to_call.invoke(tool_call["args"]) # tool_outputs.append({"output": tool_output, "tool_call_id": tool_call["id"]}) # return {"intermediate_steps": tool_outputs} # # This setup is more complex but correctly models the ReAct loop in LangGraph # class BasicAgent: # def __init__(self): # if not hf_token or not tavily_api_key: # raise ValueError("HF_TOKEN or TAVILY_API_KEY not set. Please add them to your Space secrets.") # print("LangGraph Agent initialized successfully.") # # We need an agent executor to run the loop # from langchain.agents import AgentExecutor # self.agent_executor = AgentExecutor(agent=agent_runnable, tools=tools, verbose=True) # def __call__(self, question: str) -> str: # print(f"Agent received question (first 80 chars): {question[:80]}...") # try: # # The AgentExecutor expects a dictionary with an "input" key. # response = self.agent_executor.invoke({"input": question}) # final_answer = response.get("output", "Agent did not produce an output.") # print(f"Agent returning final answer (first 80 chars): {final_answer[:80]}...") # return final_answer # except Exception as e: # print(f"An error occurred in agent execution: {e}") # return f"Error: {e}" # # ============================================================================== # # PART 2: THE GRADIO TEST HARNESS UI (UNCHANGED) # # ============================================================================== # # --- Constants --- # DEFAULT_API_URL = "https://agents-course-unit4-scoring.hf.space" # def run_and_submit_all(profile: gr.OAuthProfile | None): # # This entire function remains the same as the template # space_id = os.getenv("SPACE_ID") # if profile: # username= f"{profile.username}" # print(f"User logged in: {username}") # else: # print("User not logged in.") # return "Please Login to Hugging Face with the button.", None # api_url = DEFAULT_API_URL # questions_url = f"{api_url}/questions" # submit_url = f"{api_url}/submit" # try: # agent = BasicAgent() # except Exception as e: # print(f"Error instantiating agent: {e}") # return f"Error initializing agent: {e}", None # agent_code = f"https://huggingface.co/spaces/{space_id}/tree/main" # print(f"Fetching questions from: {questions_url}") # try: # response = requests.get(questions_url, timeout=15) # response.raise_for_status() # questions_data = response.json() # print(f"Fetched {len(questions_data)} questions.") # except Exception as e: # return f"An unexpected error occurred fetching questions: {e}", None # results_log, answers_payload = [], [] # print(f"Running agent on {len(questions_data)} questions...") # for item in questions_data: # task_id, question_text = item.get("task_id"), item.get("question") # if not task_id or question_text is None: continue # try: # submitted_answer = agent(question_text) # answers_payload.append({"task_id": task_id, "submitted_answer": submitted_answer}) # results_log.append({"Task ID": task_id, "Question": question_text, "Submitted Answer": submitted_answer}) # except Exception as e: # results_log.append({"Task ID": task_id, "Question": question_text, "Submitted Answer": f"AGENT ERROR: {e}"}) # if not answers_payload: return "Agent did not produce any answers to submit.", pd.DataFrame(results_log) # submission_data = {"username": username.strip(), "agent_code": agent_code, "answers": answers_payload} # print(f"Submitting {len(answers_payload)} answers to: {submit_url}") # try: # response = requests.post(submit_url, json=submission_data, timeout=60) # response.raise_for_status() # result_data = response.json() # final_status = (f"Submission Successful!\nUser: {result_data.get('username')}\nOverall Score: {result_data.get('score', 'N/A')}% ({result_data.get('correct_count', '?')}/{result_data.get('total_attempted', '?')} correct)\nMessage: {result_data.get('message', '')}") # return final_status, pd.DataFrame(results_log) # except Exception as e: # return f"An unexpected error occurred during submission: {e}", pd.DataFrame(results_log) # # --- Gradio Interface (Unchanged) --- # with gr.Blocks() as demo: # gr.Markdown("# GAIA Agent Evaluation Runner") # gr.Markdown("1. Log in. 2. Click 'Run Evaluation'.") # gr.LoginButton() # run_button = gr.Button("Run Evaluation & Submit All Answers") # status_output = gr.Textbox(label="Run Status / Submission Result", lines=5, interactive=False) # results_table = gr.DataFrame(label="Questions and Agent Answers", wrap=True) # run_button.click(fn=run_and_submit_all, outputs=[status_output, results_table]) # if __name__ == "__main__": # demo.launch(debug=True, share=False)