new smolagent

app.py

@@ -1,215 +1,367 @@
-"""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
+import gradio as gr
+import requests
+import inspect
+import pandas as pd
+from core_agent import GAIAAgent
 
-@tool
-def add(a: int, b: int) -> int:
-    """Add two numbers.
+# Debug function to show available environment variables
+def debug_environment():
+    """Print available environment variables related to API keys (with values hidden)"""
+    debug_vars = [
+        "HF_API_TOKEN", "HUGGINGFACEHUB_API_TOKEN", 
+        "OPENAI_API_KEY", "XAI_API_KEY",
+        "AGENT_MODEL_TYPE", "AGENT_MODEL_ID", 
+        "AGENT_TEMPERATURE", "AGENT_VERBOSE"
+    ]
     
-    Args:
-        a: first int
-        b: second int
-    """
-    return a + b
+    print("=== DEBUG: Environment Variables ===")
+    for var in debug_vars:
+        if os.environ.get(var):
+            # Hide actual values for security
+            print(f"{var}: [SET]")
+        else:
+            print(f"{var}: [NOT SET]")
+    print("===================================")
 
-@tool
-def subtract(a: int, b: int) -> int:
-    """Subtract two numbers.
-    
-    Args:
-        a: first int
-        b: second int
-    """
-    return a - b
+# (Keep Constants as is)
+# --- Constants ---
+DEFAULT_API_URL = "https://agents-course-unit4-scoring.hf.space"
 
-@tool
-def divide(a: int, b: int) -> int:
-    """Divide two numbers.
-    
-    Args:
-        a: first int
-        b: second int
-    """
-    if b == 0:
-        raise ValueError("Cannot divide by zero.")
-    return a / b
+# --- Basic Agent Definition ---
+# ----- THIS IS WERE YOU CAN BUILD WHAT YOU WANT ------
+class BasicAgent:
+    def __init__(self):
+        print("BasicAgent initialized.")
+        
+        # Call debug function to show available environment variables
+        debug_environment()
+        
+        # Initialize the GAIAAgent with local execution
+        try:
+            # Load environment variables if dotenv is available
+            try:
+                import dotenv
+                dotenv.load_dotenv()
+                print("Loaded environment variables from .env file")
+            except ImportError:
+                print("python-dotenv not installed, continuing with environment as is")
+            
+            # Try to load API keys from environment
+            # Check both HF_API_TOKEN and HUGGINGFACEHUB_API_TOKEN (HF Spaces uses HF_API_TOKEN)
+            hf_token = os.environ.get("HF_API_TOKEN") or os.environ.get("HUGGINGFACEHUB_API_TOKEN")
+            openai_key = os.environ.get("OPENAI_API_KEY")
+            xai_key = os.environ.get("XAI_API_KEY")
+            
+            # If we have at least one API key, use a model-based approach
+            if hf_token or openai_key or xai_key:
+                # Default model parameters - read directly from environment
+                model_type = os.environ.get("AGENT_MODEL_TYPE", "OpenAIServerModel")
+                model_id = os.environ.get("AGENT_MODEL_ID", "gpt-4o")
+                temperature = float(os.environ.get("AGENT_TEMPERATURE", "0.2"))
+                verbose = os.environ.get("AGENT_VERBOSE", "false").lower() == "true"
+                
+                print(f"Agent config - Model Type: {model_type}, Model ID: {model_id}")
+                
+                try:
+                    if xai_key:
+                        # Use X.AI API with OpenAIServerModel
+                        api_base = os.environ.get("XAI_API_BASE", "https://api.x.ai/v1")
+                        self.gaia_agent = GAIAAgent(
+                            model_type="OpenAIServerModel",
+                            model_id="grok-3-latest",  # X.AI's model
+                            api_key=xai_key,
+                            api_base=api_base,
+                            temperature=temperature,
+                            executor_type="local",
+                            verbose=verbose
+                        )
+                        print(f"Using OpenAIServerModel with X.AI API at {api_base}")
+                    elif model_type == "HfApiModel" and hf_token:
+                        # Use Hugging Face API
+                        self.gaia_agent = GAIAAgent(
+                            model_type="HfApiModel",
+                            model_id=model_id,
+                            api_key=hf_token,
+                            temperature=temperature,
+                            executor_type="local",
+                            verbose=verbose
+                        )
+                        print(f"Using HfApiModel with model_id: {model_id}")
+                    elif openai_key:
+                        # Default to OpenAI API
+                        api_base = os.environ.get("AGENT_API_BASE")
+                        kwargs = {
+                            "model_type": "OpenAIServerModel",
+                            "model_id": model_id,
+                            "api_key": openai_key,
+                            "temperature": temperature,
+                            "executor_type": "local",
+                            "verbose": verbose
+                        }
+                        if api_base:
+                            kwargs["api_base"] = api_base
+                            print(f"Using custom API base: {api_base}")
+                        
+                        self.gaia_agent = GAIAAgent(**kwargs)
+                        print(f"Using OpenAIServerModel with model_id: {model_id}")
+                    else:
+                        # Fallback to using whatever token we have
+                        print("WARNING: Using fallback initialization with available token")
+                        if hf_token:
+                            self.gaia_agent = GAIAAgent(
+                                model_type="HfApiModel",
+                                model_id="mistralai/Mistral-7B-Instruct-v0.2",
+                                api_key=hf_token,
+                                temperature=temperature,
+                                executor_type="local",
+                                verbose=verbose
+                            )
+                        elif openai_key:
+                            self.gaia_agent = GAIAAgent(
+                                model_type="OpenAIServerModel",
+                                model_id="gpt-3.5-turbo",
+                                api_key=openai_key,
+                                temperature=temperature,
+                                executor_type="local",
+                                verbose=verbose
+                            )
+                        else:
+                            self.gaia_agent = GAIAAgent(
+                                model_type="OpenAIServerModel",
+                                model_id="grok-3-latest",
+                                api_key=xai_key,
+                                api_base=os.environ.get("XAI_API_BASE", "https://api.x.ai/v1"),
+                                temperature=temperature,
+                                executor_type="local",
+                                verbose=verbose
+                            )
+                except ImportError as ie:
+                    # Handle OpenAI module errors specifically
+                    if "openai" in str(ie).lower() and hf_token:
+                        print(f"OpenAI module error: {ie}. Falling back to HfApiModel.")
+                        self.gaia_agent = GAIAAgent(
+                            model_type="HfApiModel",
+                            model_id="mistralai/Mistral-7B-Instruct-v0.2",
+                            api_key=hf_token,
+                            temperature=temperature,
+                            executor_type="local",
+                            verbose=verbose
+                        )
+                        print(f"Using HfApiModel with model_id: mistralai/Mistral-7B-Instruct-v0.2 (fallback)")
+                    else:
+                        raise
+            else:
+                # No API keys available, log the error
+                print("ERROR: No API keys found. Please set at least one of these environment variables:")
+                print("- HUGGINGFACEHUB_API_TOKEN or HF_API_TOKEN")
+                print("- OPENAI_API_KEY")
+                print("- XAI_API_KEY")
+                self.gaia_agent = None
+                print("WARNING: No API keys found. Agent will not be able to answer questions.")
+                
+        except Exception as e:
+            print(f"Error initializing GAIAAgent: {e}")
+            self.gaia_agent = None
+            print("WARNING: Failed to initialize agent. Falling back to basic responses.")
+            
+    def __call__(self, question: str) -> str:
+        print(f"Agent received question (first 50 chars): {question[:50]}...")
+        
+        # Check if we have a functioning GAIA agent
+        if self.gaia_agent:
+            try:
+                # Process the question using the GAIA agent
+                answer = self.gaia_agent.answer_question(question)
+                print(f"Agent generated answer: {answer[:50]}..." if len(answer) > 50 else f"Agent generated answer: {answer}")
+                return answer
+            except Exception as e:
+                print(f"Error processing question: {e}")
+                # Fall back to a simple response on error
+                return "An error occurred while processing your question. Please check the agent logs for details."
+        else:
+            # We don't have a valid agent, provide a basic response
+            return "The agent is not properly initialized. Please check your API keys and configuration."
 
-@tool
-def modulus(a: int, b: int) -> int:
-    """Get the modulus of two numbers.
-    
-    Args:
-        a: first int
-        b: second int
+def run_and_submit_all( profile: gr.OAuthProfile | None):
     """
-    return a % b
+    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
 
-@tool
-def wiki_search(query: str) -> str:
-    """Search Wikipedia for a query and return maximum 2 results.
-    
-    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.
-    
-    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.
-    
-    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}
-
-
-
-# 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,
-            ),
-        )
+    if profile:
+        username= f"{profile.username}"
+        print(f"User logged in: {username}")
     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 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}",
+        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()
+        
+        # Check if agent is properly initialized
+        if not agent.gaia_agent:
+            print("ERROR: Agent was not properly initialized")
+            return "ERROR: Agent was not properly initialized. Check the logs for details on missing API keys or configuration.", None
+        
+    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.')}"
         )
-        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,
+        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.
+        """
     )
-    builder.add_edge("tools", "assistant")
 
-    # Compile graph
-    return builder.compile()
+    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]
+    )
 
-# 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()
+    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)
 

core_agent.py

@@ -0,0 +1,493 @@
+from smolagents import (
+    CodeAgent, 
+    DuckDuckGoSearchTool, 
+    HfApiModel, 
+    LiteLLMModel,
+    OpenAIServerModel,
+    PythonInterpreterTool,
+    tool,
+    InferenceClientModel
+)
+from typing import List, Dict, Any, Optional
+import os
+import tempfile
+import re
+import json
+import requests
+from urllib.parse import urlparse
+
+@tool
+def save_and_read_file(content: str, filename: Optional[str] = None) -> str:
+    """
+    Save content to a temporary file and return the path.
+    Useful for processing files from the GAIA API.
+    
+    Args:
+        content: The content to save to the file
+        filename: Optional filename, will generate a random name if not provided
+        
+    Returns:
+        Path to the saved file
+    """
+    temp_dir = tempfile.gettempdir()
+    if filename is None:
+        temp_file = tempfile.NamedTemporaryFile(delete=False)
+        filepath = temp_file.name
+    else:
+        filepath = os.path.join(temp_dir, filename)
+    
+    # Write content to the file
+    with open(filepath, 'w') as f:
+        f.write(content)
+    
+    return f"File saved to {filepath}. You can read this file to process its contents."
+
+@tool
+def download_file_from_url(url: str, filename: Optional[str] = None) -> str:
+    """
+    Download a file from a URL and save it to a temporary location.
+    
+    Args:
+        url: The URL to download from
+        filename: Optional filename, will generate one based on URL if not provided
+        
+    Returns:
+        Path to the downloaded file
+    """
+    try:
+        # Parse URL to get filename if not provided
+        if not filename:
+            path = urlparse(url).path
+            filename = os.path.basename(path)
+            if not filename:
+                # Generate a random name if we couldn't extract one
+                import uuid
+                filename = f"downloaded_{uuid.uuid4().hex[:8]}"
+        
+        # Create temporary file
+        temp_dir = tempfile.gettempdir()
+        filepath = os.path.join(temp_dir, filename)
+        
+        # Download the file
+        response = requests.get(url, stream=True)
+        response.raise_for_status()
+        
+        # Save the file
+        with open(filepath, 'wb') as f:
+            for chunk in response.iter_content(chunk_size=8192):
+                f.write(chunk)
+        
+        return f"File downloaded to {filepath}. You can now process this file."
+    except Exception as e:
+        return f"Error downloading file: {str(e)}"
+
+@tool
+def extract_text_from_image(image_path: str) -> str:
+    """
+    Extract text from an image using pytesseract (if available).
+    
+    Args:
+        image_path: Path to the image file
+        
+    Returns:
+        Extracted text or error message
+    """
+    try:
+        # Try to import pytesseract
+        import pytesseract
+        from PIL import Image
+        
+        # Open the image
+        image = Image.open(image_path)
+        
+        # Extract text
+        text = pytesseract.image_to_string(image)
+        
+        return f"Extracted text from image:\n\n{text}"
+    except ImportError:
+        return "Error: pytesseract is not installed. Please install it with 'pip install pytesseract' and ensure Tesseract OCR is installed on your system."
+    except Exception as e:
+        return f"Error extracting text from image: {str(e)}"
+
+@tool
+def analyze_csv_file(file_path: str, query: str) -> str:
+    """
+    Analyze a CSV file using pandas and answer a question about it.
+    
+    Args:
+        file_path: Path to the CSV file
+        query: Question about the data
+        
+    Returns:
+        Analysis result or error message
+    """
+    try:
+        import pandas as pd
+        
+        # Read the CSV file
+        df = pd.read_csv(file_path)
+        
+        # Run various analyses based on the query
+        result = f"CSV file loaded with {len(df)} rows and {len(df.columns)} columns.\n"
+        result += f"Columns: {', '.join(df.columns)}\n\n"
+        
+        # Add summary statistics
+        result += "Summary statistics:\n"
+        result += str(df.describe())
+        
+        return result
+    except ImportError:
+        return "Error: pandas is not installed. Please install it with 'pip install pandas'."
+    except Exception as e:
+        return f"Error analyzing CSV file: {str(e)}"
+
+@tool
+def analyze_excel_file(file_path: str, query: str) -> str:
+    """
+    Analyze an Excel file using pandas and answer a question about it.
+    
+    Args:
+        file_path: Path to the Excel file
+        query: Question about the data
+        
+    Returns:
+        Analysis result or error message
+    """
+    try:
+        import pandas as pd
+        
+        # Read the Excel file
+        df = pd.read_excel(file_path)
+        
+        # Run various analyses based on the query
+        result = f"Excel file loaded with {len(df)} rows and {len(df.columns)} columns.\n"
+        result += f"Columns: {', '.join(df.columns)}\n\n"
+        
+        # Add summary statistics
+        result += "Summary statistics:\n"
+        result += str(df.describe())
+        
+        return result
+    except ImportError:
+        return "Error: pandas and openpyxl are not installed. Please install them with 'pip install pandas openpyxl'."
+    except Exception as e:
+        return f"Error analyzing Excel file: {str(e)}"
+
+class GAIAAgent:
+    def __init__(
+        self, 
+        model_type: str = "HfApiModel", 
+        model_id: Optional[str] = None,
+        api_key: Optional[str] = None,
+        api_base: Optional[str] = None,
+        temperature: float = 0.2,
+        executor_type: str = "local",  # Changed from use_e2b to executor_type
+        additional_imports: List[str] = None,
+        additional_tools: List[Any] = None,
+        system_prompt: Optional[str] = None,  # We'll still accept this parameter but not use it directly
+        verbose: bool = False,
+        provider: Optional[str] = None,  # Add provider for InferenceClientModel
+        timeout: Optional[int] = None   # Add timeout for InferenceClientModel
+    ):
+        """
+        Initialize a GAIAAgent with specified configuration
+        
+        Args:
+            model_type: Type of model to use (HfApiModel, LiteLLMModel, OpenAIServerModel, InferenceClientModel)
+            model_id: ID of the model to use
+            api_key: API key for the model provider
+            api_base: Base URL for API calls
+            temperature: Temperature for text generation
+            executor_type: Type of executor for code execution ('local' or 'e2b')
+            additional_imports: Additional Python modules to allow importing
+            additional_tools: Additional tools to provide to the agent
+            system_prompt: Custom system prompt to use (not directly used, kept for backward compatibility)
+            verbose: Enable verbose logging
+            provider: Provider for InferenceClientModel (e.g., "hf-inference")
+            timeout: Timeout in seconds for API calls
+        """
+        # Set verbosity
+        self.verbose = verbose
+        self.system_prompt = system_prompt  # Store for potential future use
+        
+        # Initialize model based on configuration
+        if model_type == "HfApiModel":
+            if api_key is None:
+                api_key = os.getenv("HUGGINGFACEHUB_API_TOKEN")
+                if not api_key:
+                    raise ValueError("No Hugging Face token provided. Please set HUGGINGFACEHUB_API_TOKEN environment variable or pass api_key parameter.")
+            
+            if self.verbose:
+                print(f"Using Hugging Face token: {api_key[:5]}...")
+                
+            self.model = HfApiModel(
+                model_id=model_id or "meta-llama/Llama-3-70B-Instruct",
+                token=api_key,
+                temperature=temperature
+            )
+        elif model_type == "InferenceClientModel":
+            if api_key is None:
+                api_key = os.getenv("HUGGINGFACEHUB_API_TOKEN")
+                if not api_key:
+                    raise ValueError("No Hugging Face token provided. Please set HUGGINGFACEHUB_API_TOKEN environment variable or pass api_key parameter.")
+            
+            if self.verbose:
+                print(f"Using Hugging Face token: {api_key[:5]}...")
+                
+            self.model = InferenceClientModel(
+                model_id=model_id or "meta-llama/Llama-3-70B-Instruct",
+                provider=provider or "hf-inference",
+                token=api_key,
+                timeout=timeout or 120,
+                temperature=temperature
+            )
+        elif model_type == "LiteLLMModel":
+            from smolagents import LiteLLMModel
+            self.model = LiteLLMModel(
+                model_id=model_id or "gpt-4o",
+                api_key=api_key or os.getenv("OPENAI_API_KEY"),
+                temperature=temperature
+            )
+        elif model_type == "OpenAIServerModel":
+            # Check for xAI API key and base URL first
+            xai_api_key = os.getenv("XAI_API_KEY")
+            xai_api_base = os.getenv("XAI_API_BASE")
+            
+            # If xAI credentials are available, use them
+            if xai_api_key and api_key is None:
+                api_key = xai_api_key
+                if self.verbose:
+                    print(f"Using xAI API key: {api_key[:5]}...")
+            
+            # If no API key specified, fall back to OPENAI_API_KEY
+            if api_key is None:
+                api_key = os.getenv("OPENAI_API_KEY")
+                if not api_key:
+                    raise ValueError("No OpenAI API key provided. Please set OPENAI_API_KEY or XAI_API_KEY environment variable or pass api_key parameter.")
+            
+            # If xAI API base is available and no api_base is provided, use it
+            if xai_api_base and api_base is None:
+                api_base = xai_api_base
+                if self.verbose:
+                    print(f"Using xAI API base URL: {api_base}")
+            
+            # If no API base specified but environment variable available, use it
+            if api_base is None:
+                api_base = os.getenv("AGENT_API_BASE")
+                if api_base and self.verbose:
+                    print(f"Using API base from AGENT_API_BASE: {api_base}")
+            
+            self.model = OpenAIServerModel(
+                model_id=model_id or "gpt-4o",
+                api_key=api_key,
+                api_base=api_base,
+                temperature=temperature
+            )
+        else:
+            raise ValueError(f"Unknown model type: {model_type}")
+        
+        if self.verbose:
+            print(f"Initialized model: {model_type} - {model_id}")
+        
+        # Initialize default tools
+        self.tools = [
+            DuckDuckGoSearchTool(),
+            PythonInterpreterTool(),
+            save_and_read_file,
+            download_file_from_url,
+            analyze_csv_file,
+            analyze_excel_file
+        ]
+        
+        # Add extract_text_from_image if PIL and pytesseract are available
+        try:
+            import pytesseract
+            from PIL import Image
+            self.tools.append(extract_text_from_image)
+            if self.verbose:
+                print("Added image processing tool")
+        except ImportError:
+            if self.verbose:
+                print("Image processing libraries not available")
+        
+        # Add any additional tools
+        if additional_tools:
+            self.tools.extend(additional_tools)
+            
+        if self.verbose:
+            print(f"Initialized with {len(self.tools)} tools")
+        
+        # Setup imports allowed
+        self.imports = ["pandas", "numpy", "datetime", "json", "re", "math", "os", "requests", "csv", "urllib"]
+        if additional_imports:
+            self.imports.extend(additional_imports)
+            
+        # Initialize the CodeAgent
+        executor_kwargs = {}
+        if executor_type == "e2b":
+            try:
+                # Try to import e2b dependencies to check if they're available
+                from e2b_code_interpreter import Sandbox
+                if self.verbose:
+                    print("Using e2b executor")
+            except ImportError:
+                if self.verbose:
+                    print("e2b dependencies not found, falling back to local executor")
+                executor_type = "local"  # Fallback to local if e2b is not available
+        
+        self.agent = CodeAgent(
+            tools=self.tools,
+            model=self.model,
+            additional_authorized_imports=self.imports,
+            executor_type=executor_type,
+            executor_kwargs=executor_kwargs,
+            verbosity_level=2 if self.verbose else 0
+        )
+        
+        if self.verbose:
+            print("Agent initialized and ready")
+    
+    def answer_question(self, question: str, task_file_path: Optional[str] = None) -> str:
+        """
+        Process a GAIA benchmark question and return the answer
+        
+        Args:
+            question: The question to answer
+            task_file_path: Optional path to a file associated with the question
+            
+        Returns:
+            The answer to the question
+        """
+        try:
+            if self.verbose:
+                print(f"Processing question: {question}")
+                if task_file_path:
+                    print(f"With associated file: {task_file_path}")
+            
+            # Create a context with file information if available
+            context = question
+            file_content = None
+            
+            # If there's a file, read it and include its content in the context
+            if task_file_path:
+                try:
+                    with open(task_file_path, 'r') as f:
+                        file_content = f.read()
+                    
+                    # Determine file type from extension
+                    import os
+                    file_ext = os.path.splitext(task_file_path)[1].lower()
+                    
+                    context = f"""
+Question: {question}
+
+This question has an associated file. Here is the file content:
+
+```{file_ext}
+{file_content}
+```
+
+Analyze the file content above to answer the question.
+"""
+                except Exception as file_e:
+                    context = f"""
+Question: {question}
+
+This question has an associated file at path: {task_file_path}
+However, there was an error reading the file: {file_e}
+You can still try to answer the question based on the information provided.
+"""
+            
+            # Check for special cases that need specific formatting
+            # Reversed text questions
+            if question.startswith(".") or ".rewsna eht sa" in question:
+                context = f"""
+This question appears to be in reversed text. Here's the reversed version:
+{question[::-1]}
+
+Now answer the question above. Remember to format your answer exactly as requested.
+"""
+            
+            # Add a prompt to ensure precise answers
+            full_prompt = f"""{context}
+
+When answering, provide ONLY the precise answer requested. 
+Do not include explanations, steps, reasoning, or additional text.
+Be direct and specific. GAIA benchmark requires exact matching answers.
+For example, if asked "What is the capital of France?", respond simply with "Paris".
+"""
+            
+            # Run the agent with the question
+            answer = self.agent.run(full_prompt)
+            
+            # Clean up the answer to ensure it's in the expected format
+            # Remove common prefixes that models often add
+            answer = self._clean_answer(answer)
+            
+            if self.verbose:
+                print(f"Generated answer: {answer}")
+                
+            return answer
+        except Exception as e:
+            error_msg = f"Error answering question: {e}"
+            if self.verbose:
+                print(error_msg)
+            return error_msg
+    
+    def _clean_answer(self, answer: any) -> str:
+        """
+        Clean up the answer to remove common prefixes and formatting
+        that models often add but that can cause exact match failures.
+        
+        Args:
+            answer: The raw answer from the model
+            
+        Returns:
+            The cleaned answer as a string
+        """
+        # Convert non-string types to strings
+        if not isinstance(answer, str):
+            # Handle numeric types (float, int)
+            if isinstance(answer, float):
+                # Format floating point numbers properly
+                # Check if it's an integer value in float form (e.g., 12.0)
+                if answer.is_integer():
+                    formatted_answer = str(int(answer))
+                else:
+                    # For currency values that might need formatting
+                    if abs(answer) >= 1000:
+                        formatted_answer = f"${answer:,.2f}"
+                    else:
+                        formatted_answer = str(answer)
+                return formatted_answer
+            elif isinstance(answer, int):
+                return str(answer)
+            else:
+                # For any other type
+                return str(answer)
+        
+        # Now we know answer is a string, so we can safely use string methods
+        # Normalize whitespace
+        answer = answer.strip()
+        
+        # Remove common prefixes and formatting that models add
+        prefixes_to_remove = [
+            "The answer is ", 
+            "Answer: ",
+            "Final answer: ",
+            "The result is ",
+            "To answer this question: ",
+            "Based on the information provided, ",
+            "According to the information: ",
+        ]
+        
+        for prefix in prefixes_to_remove:
+            if answer.startswith(prefix):
+                answer = answer[len(prefix):].strip()
+        
+        # Remove quotes if they wrap the entire answer
+        if (answer.startswith('"') and answer.endswith('"')) or (answer.startswith("'") and answer.endswith("'")):
+            answer = answer[1:-1].strip()
+        
+        return answer
+

main.py

@@ -0,0 +1,278 @@
+import os
+import tempfile
+import gradio as gr
+import pandas as pd
+import traceback
+from core_agent import GAIAAgent
+from api_integration import GAIAApiClient
+
+# Constants
+DEFAULT_API_URL = "https://agents-course-unit4-scoring.hf.space"
+
+def save_task_file(file_content, task_id):
+    """
+    Save a task file to a temporary location
+    """
+    if not file_content:
+        return None
+    
+    # Create a temporary file
+    temp_dir = tempfile.gettempdir()
+    file_path = os.path.join(temp_dir, f"gaia_task_{task_id}.txt")
+    
+    # Write content to the file
+    with open(file_path, 'wb') as f:
+        f.write(file_content)
+    
+    print(f"File saved to {file_path}")
+    return file_path
+
+def get_agent_configuration():
+    """
+    Get the agent configuration based on environment variables
+    """
+    # Default configuration
+    config = {
+        "model_type": "OpenAIServerModel",  # Default to OpenAIServerModel
+        "model_id": "gpt-4o",  # Default model for OpenAI
+        "temperature": 0.2,
+        "executor_type": "local",
+        "verbose": False,
+        "provider": "hf-inference",  # For InferenceClientModel
+        "timeout": 120        # For InferenceClientModel
+    }
+    
+    # Check for xAI API key and base URL
+    xai_api_key = os.getenv("XAI_API_KEY")
+    xai_api_base = os.getenv("XAI_API_BASE")
+    
+    # If we have xAI credentials, use them
+    if xai_api_key:
+        config["api_key"] = xai_api_key
+        if xai_api_base:
+            config["api_base"] = xai_api_base
+            # Use a model that works well with xAI
+            config["model_id"] = "mixtral-8x7b-32768"
+    
+    # Override with environment variables if present
+    if os.getenv("AGENT_MODEL_TYPE"):
+        config["model_type"] = os.getenv("AGENT_MODEL_TYPE")
+    
+    if os.getenv("AGENT_MODEL_ID"):
+        config["model_id"] = os.getenv("AGENT_MODEL_ID")
+    
+    if os.getenv("AGENT_TEMPERATURE"):
+        config["temperature"] = float(os.getenv("AGENT_TEMPERATURE"))
+    
+    if os.getenv("AGENT_EXECUTOR_TYPE"):
+        config["executor_type"] = os.getenv("AGENT_EXECUTOR_TYPE")
+    
+    if os.getenv("AGENT_VERBOSE") is not None:
+        config["verbose"] = os.getenv("AGENT_VERBOSE").lower() == "true"
+    
+    if os.getenv("AGENT_API_BASE"):
+        config["api_base"] = os.getenv("AGENT_API_BASE")
+    
+    # InferenceClientModel specific settings
+    if os.getenv("AGENT_PROVIDER"):
+        config["provider"] = os.getenv("AGENT_PROVIDER")
+    
+    if os.getenv("AGENT_TIMEOUT"):
+        config["timeout"] = int(os.getenv("AGENT_TIMEOUT"))
+    
+    return config
+
+def run_and_submit_all(profile: gr.OAuthProfile | None):
+    """
+    Fetches all questions, runs the GAIAAgent on them, submits all answers,
+    and displays the results.
+    """
+    # Check for user login
+    if not profile:
+        return "Please Login to Hugging Face with the button.", None
+    
+    username = profile.username
+    print(f"User logged in: {username}")
+    
+    # Get SPACE_ID for code link
+    space_id = os.getenv("SPACE_ID")
+    agent_code = f"https://huggingface.co/spaces/{space_id}/tree/main"
+    
+    # Initialize API client
+    api_client = GAIAApiClient(DEFAULT_API_URL)
+    
+    # Initialize Agent with configuration
+    try:
+        agent_config = get_agent_configuration()
+        print(f"Using agent configuration: {agent_config}")
+        
+        agent = GAIAAgent(**agent_config)
+        print("Agent initialized successfully")
+    except Exception as e:
+        error_details = traceback.format_exc()
+        print(f"Error initializing agent: {e}\n{error_details}")
+        return f"Error initializing agent: {e}", None
+    
+    # Fetch questions
+    try:
+        questions_data = api_client.get_questions()
+        if not questions_data:
+            return "Fetched questions list is empty or invalid format.", None
+        print(f"Fetched {len(questions_data)} questions.")
+    except Exception as e:
+        error_details = traceback.format_exc()
+        print(f"Error fetching questions: {e}\n{error_details}")
+        return f"Error fetching questions: {e}", None
+    
+    # Run agent on questions
+    results_log = []
+    answers_payload = []
+    print(f"Running agent on {len(questions_data)} questions...")
+    
+    # Progress tracking
+    total_questions = len(questions_data)
+    completed = 0
+    failed = 0
+    
+    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:
+            # Update progress
+            completed += 1
+            print(f"Processing question {completed}/{total_questions}: Task ID {task_id}")
+            
+            # Check if the question has an associated file
+            file_path = None
+            try:
+                file_content = api_client.get_file(task_id)
+                print(f"Downloaded file for task {task_id}")
+                file_path = save_task_file(file_content, task_id)
+            except Exception as file_e:
+                print(f"No file found for task {task_id} or error: {file_e}")
+            
+            # Run the agent to get the answer
+            submitted_answer = agent.answer_question(question_text, file_path)
+            
+            # Add to results
+            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:
+            # Update error count
+            failed += 1
+            error_details = traceback.format_exc()
+            print(f"Error running agent on task {task_id}: {e}\n{error_details}")
+            
+            # Add error to results
+            error_msg = f"AGENT ERROR: {e}"
+            answers_payload.append({"task_id": task_id, "submitted_answer": error_msg})
+            results_log.append({
+                "Task ID": task_id,
+                "Question": question_text,
+                "Submitted Answer": error_msg
+            })
+    
+    # Print summary
+    print(f"\nProcessing complete: {completed} questions processed, {failed} failures")
+    
+    if not answers_payload:
+        return "Agent did not produce any answers to submit.", pd.DataFrame(results_log)
+    
+    # Submit answers
+    submission_data = {
+        "username": username.strip(),
+        "agent_code": agent_code,
+        "answers": answers_payload
+    }
+    
+    print(f"Submitting {len(answers_payload)} answers for username '{username}'...")
+    
+    try:
+        result_data = api_client.submit_answers(
+            username.strip(),
+            agent_code,
+            answers_payload
+        )
+        
+        # Calculate success rate
+        correct_count = result_data.get('correct_count', 0)
+        total_attempted = result_data.get('total_attempted', len(answers_payload))
+        success_rate = (correct_count / total_attempted) * 100 if total_attempted > 0 else 0
+        
+        final_status = (
+            f"Submission Successful!\n"
+            f"User: {result_data.get('username')}\n"
+            f"Overall Score: {result_data.get('score', 'N/A')}% "
+            f"({correct_count}/{total_attempted} correct, {success_rate:.1f}% success rate)\n"
+            f"Message: {result_data.get('message', 'No message received.')}"
+        )
+        
+        print("Submission successful.")
+        return final_status, pd.DataFrame(results_log)
+    except Exception as e:
+        error_details = traceback.format_exc()
+        status_message = f"Submission Failed: {e}\n{error_details}"
+        print(status_message)
+        return status_message, pd.DataFrame(results_log)
+
+# Build Gradio Interface
+with gr.Blocks() as demo:
+    gr.Markdown("# GAIA Agent Evaluation Runner")
+    gr.Markdown(
+        """
+        **Instructions:**
+        
+        1. Log in to your Hugging Face account using the button below.
+        2. Click 'Run Evaluation & Submit All Answers' to fetch questions, run your agent, submit answers, and see the score.
+        
+        **Configuration:**
+        
+        You can configure the agent by setting these environment variables:
+        - `AGENT_MODEL_TYPE`: Model type (HfApiModel, InferenceClientModel, LiteLLMModel, OpenAIServerModel)
+        - `AGENT_MODEL_ID`: Model ID
+        - `AGENT_TEMPERATURE`: Temperature for generation (0.0-1.0)
+        - `AGENT_EXECUTOR_TYPE`: Type of executor ('local' or 'e2b')
+        - `AGENT_VERBOSE`: Enable verbose logging (true/false)
+        - `AGENT_API_BASE`: Base URL for API calls (for OpenAIServerModel)
+        
+        **xAI Support:**
+        - `XAI_API_KEY`: Your xAI API key
+        - `XAI_API_BASE`: Base URL for xAI API (default: https://api.groq.com/openai/v1)
+        - When using xAI, set AGENT_MODEL_TYPE=OpenAIServerModel and AGENT_MODEL_ID=mixtral-8x7b-32768
+        
+        **InferenceClientModel specific settings:**
+        - `AGENT_PROVIDER`: Provider for InferenceClientModel (e.g., "hf-inference") 
+        - `AGENT_TIMEOUT`: Timeout in seconds for API calls
+        """
+    )
+    
+    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__":
+    print("\n" + "-"*30 + " App Starting " + "-"*30)
+    
+    # Check for environment variables
+    config = get_agent_configuration()
+    print(f"Agent configuration: {config}")
+    
+    # Run the Gradio app
+    demo.launch(debug=True, share=False)
+

requirements.txt

@@ -1,19 +1,8 @@
 gradio
 requests
-langchain
-langchain-community
-langchain-core
-langchain-google-genai
-langchain-huggingface
-langchain-groq
-langchain-tavily
-langchain-chroma
-langgraph
-huggingface_hub
-pgvector
-supabase
-arxiv
-pymupdf
-wikipedia
+smolagents[openai]
 python-dotenv
+pandas
+numpy
+openai