app redone

agent.py

@@ -101,57 +101,62 @@ def build_graph(provider: str = "mistral"):
 
     return builder.compile()
 
+# =============================================================================
+# 3. GAIAgent WRAPPER
+# =============================================================================
+class GAIAgent:
+    def __init__(self):
+        self.graph = build_graph(provider="mistral")
+
+    def _clean_answer(self, text: str) -> str:
+        """
+        Pulisce la risposta grezza dell'LLM per l'Exact Match della leaderboard.
+        """
+        if not text: 
+            return ""
+        
+        bold_match = re.search(r"\*\*(.*?)\*\*", text)
+        if bold_match:
+            return bold_match.group(1).strip()
+
+        match = re.search(r"FINAL\s*ANSWER\s*:\s*(.*)", text, re.IGNORECASE | re.DOTALL)
+        if match: 
+            text = match.group(1).strip()
+
+        text = text.replace("**", "").replace("`", "").strip()
+        
+        lines = [L.strip() for L in text.split('\n') if L.strip()]
+        if lines: text = lines[0]
+        
+        if text.endswith(".") and len(text) > 4: 
+            text = text[:-1]
+            
+        return text
 
-# ----------------------------
-# Minimal Test Execution
-# ----------------------------
-if __name__ == "__main__":
-    gaia_tasks = [
-        {
-            "id": "6f37996b",
-            "question": """Given this table defining * on the set S = {a, b, c, d, e}
-|*|a|b|c|d|e|
-|---|---|---|---|---|---|
-|a|a|b|c|b|d|
-|b|b|c|a|e|c|
-|c|c|a|b|b|a|
-|d|b|e|b|e|d|
-|e|d|b|a|d|c|
-provide the subset of S involved in any possible counter-examples that prove * is not commutative. 
-Provide your answer as a comma separated list of the elements in the set in alphabetical order."""
-        },
-        {
-            "id": "305ac316",
-            "question": "Who did the actor who played Ray in the Polish-language version of Everybody Loves Raymond play in Magda M.? Give only the first name."
-        },
-        {
-            "id": "840bfca7",
-            "question": "On June 6, 2023, an article by Carolyn Collins Petersen was published in Universe Today. This article mentions a team that produced a paper about their observations, linked at the bottom of the article. Find this paper. Under what NASA award number was the work performed by R. G. Arendt supported by?"
-        }
-    ]
+    def __call__(self, question: str) -> str:
+        """
+        Interfaccia pubblica chiamata da app.py.
+        Accetta una stringa (domanda), restituisce una stringa (risposta pulita).
+        """
+        messages = [HumanMessage(content=question)]
+        result = self.graph.invoke({"messages": messages}, config={"recursion_limit": 60})
+        raw_answer = result["messages"][-1].content
+        
+        return self._clean_answer(raw_answer)
 
-    try:
-        graph = build_graph(provider="mistral") 
-
-        for i, task in enumerate(gaia_tasks, 1):
-            print(f"\n{'#'*60}")
-            print(f"STARTING TASK {i} [ID: {task['id']}]")
-            print(f"{'#'*60}")
-
-            try:
-                messages = [HumanMessage(content=task['question'])]
-                result = graph.invoke(
-                    {"messages": messages},
-                    config={"recursion_limit": 100}
-                )
-                
-                print("\nFULL EXECUTION LOG:")
-                for m in result["messages"]:
-                    m.pretty_print()
-                    print("-" * 40)
-
-            except Exception as e:
-                print(f"ERROR executing task {task['id']}: {e}")
 
+# =============================================================================
+# 4. MAIN TEST
+# =============================================================================
+if __name__ == "__main__":
+    test_q = "How many studio albums were published by Mercedes Sosa between 2000 and 2009 (included)? You can use the latest 2022 version of english wikipedia."
+    
+    print("--- 🚀 Starting GAIAgent ---")
+    
+    try:
+        agent = GAIAgent()
+        print(f"❓ Question: {test_q}")
+        ans = agent(test_q)
+        print(f"✅ FINAL ANSWER: {ans}")
     except Exception as e:
-        print(f"CRITICAL SETUP ERROR: {e}")
+        print(f"❌ Error: {e}")

app.py

@@ -1,265 +1,214 @@
 import os
-import re
-import gradio as gr
+import json
 import requests
-import inspect
 import pandas as pd
+import gradio as gr
+import time
 
-from langchain_core.messages import HumanMessage
-from agent import build_graph
+# Import the modular agent
+try:
+    from agent import GAIAgent
+except ImportError:
+    print("WARNING: agent.py not found.")
 
-# (Keep Constants as is)
 # --- Constants ---
 DEFAULT_API_URL = "https://agents-course-unit4-scoring.hf.space"
+CACHE_FILE = "answers_cache.json"
 
-# --- GAIAgent Definition ---
-# ----- THIS IS WERE YOU CAN BUILD WHAT YOU WANT ------
-class GAIAgent:
-    def __init__(self):
-        self.graph = build_graph(provider = "groq")
+# --- RATE LIMIT CONFIGURATION ---
+MAX_RETRIES = 3       # Max attempts per question
+REQUEST_DELAY = 5     # Seconds to wait between questions (Grace Time)
+ERROR_DELAY = 10      # Seconds to wait after an error
 
-    
-    def _clean_answer(self, text: str) -> str:
-        """
-        Pulisce la risposta per garantire l'Exact Match.
-        Strategia:
-        1. Se c'è qualcosa tra **doppi asterischi**, è probabilmente la risposta.
-        2. Se c'è 'FINAL ANSWER:', prendi quello che segue.
-        3. Pulisci markdown e punteggiatura.
-        """
-        if not text: 
-            return ""
+def fetch_questions():
+    """Fetches questions from the API."""
+    try:
+        response = requests.get(f"{DEFAULT_API_URL}/questions", timeout=15)
+        response.raise_for_status()
+        return response.json()
+    except Exception as e:
+        print(f"Error fetching questions: {e}")
+        return []
 
-        # --- STRATEGIA 1: CATTURA GRASSETTO (Priorità Alta) ---
-        # Se l'LLM scrive "The answer is **128**", prendiamo solo "128".
-        bold_match = re.search(r"\*\*(.*?)\*\*", text)
-        if bold_match:
-            return bold_match.group(1).strip()
+def load_cache_as_dict():
+    """Loads cache for fast lookup."""
+    if os.path.exists(CACHE_FILE):
+        try:
+            with open(CACHE_FILE, "r") as f:
+                data = json.load(f)
+                return {item["task_id"]: item["submitted_answer"] for item in data}
+        except:
+            pass
+    return {}
+
+def save_cache(data):
+    """Saves progress to file."""
+    try:
+        with open(CACHE_FILE, "w") as f:
+            json.dump(data, f, indent=2)
+    except Exception as e:
+        print(f"Error saving cache: {e}")
+
+# =============================================================================
+# 🔎 VALIDATION LOGIC
+# =============================================================================
+def is_valid_answer(answer: str) -> bool:
+    """Checks if the answer meets minimum criteria."""
+    if not answer: return False
+    if "AGENT ERROR" in answer or "Error:" in answer: return False
+    if len(answer) > 150: return False # Too long
         
-        # --- STRATEGIA 2: CATTURA PREFIX (Standard) ---
-        match = re.search(r"FINAL\s*ANSWER\s*:\s*(.*)", text, re.IGNORECASE | re.DOTALL)
-        if match: 
-            text = match.group(1).strip()
+    forbidden_phrases = ["I cannot", "I am sorry", "The answer is", "Based on"]
+    for phrase in forbidden_phrases:
+        if phrase.lower() in answer.lower():
+            if "not known" in answer.lower(): return True
+            return False
+    return True
+
+# =============================================================================
+# 1. GENERATOR FUNCTION (With Grace Time & Retry)
+# =============================================================================
+def generate_answers(profile: gr.OAuthProfile | None):
+    """
+    Runs the agent with strategic pauses to avoid Rate Limits.
+    """
+    if not profile:
+        yield "⚠️ Please login with Hugging Face first!", pd.DataFrame()
+        return
 
-        # --- PULIZIA GENERALE ---
-        # Rimuove markdown residuo
-        text = text.replace("**", "").replace("`", "").strip()
-        
-        # Prende solo la prima riga (spesso le spiegazioni sono sotto)
-        lines = [L.strip() for L in text.split('\n') if L.strip()]
-        if lines: text = lines[0]
+    existing_answers = load_cache_as_dict()
+    questions = fetch_questions()
+    
+    if not questions:
+        yield "❌ Failed to fetch questions.", pd.DataFrame()
+        return
 
-        # --- STRATEGIA 3: HEURISTICA "THE ANSWER IS..." ---
-        # Se dopo tutto ciò abbiamo ancora una frase lunga che finisce con il codice
-        # Esempio: "The number is 80GSFC21M0002."
-        # Proviamo a splittare su " is " o " are "
-        if len(text.split()) > 5: # Se è una frase lunga
-            split_match = re.search(r"\b(is|are|was|were)\b\s*(.*)$", text, re.IGNORECASE)
-            if split_match:
-                candidate = split_match.group(2).strip()
-                # Se quello che rimane è breve (la risposta), prendiamo quello
-                if len(candidate.split()) <= 3:
-                    text = candidate
+    try:
+        agent = GAIAgent()
+    except Exception as e:
+        yield f"❌ Error initializing agent: {e}", pd.DataFrame()
+        return
 
-        # Rimuove punto finale
-        if text.endswith(".") and len(text) > 4: 
-            text = text[:-1]
-            
-        return text
+    results_log = []     
+    payload_cache = []   
+    total = len(questions)
     
+    yield f"🚀 Starting with {REQUEST_DELAY}s grace time...", pd.DataFrame()
 
-    def __call__(self, question: str) -> str:
-        from langchain_core.messages import HumanMessage
-        messages = [HumanMessage(content=question)]
+    for i, item in enumerate(questions):
+        task_id = item.get("task_id")
+        question_text = item.get("question")
         
-        # Eseguiamo il grafo con un limite di ricorsione alto
-        result = self.graph.invoke(
-            {"messages": messages},
-            config={"recursion_limit": 100} 
-        )
+        # --- CACHE CHECK (Fast) ---
+        if task_id in existing_answers:
+            answer = existing_answers[task_id]
+            if is_valid_answer(answer):
+                results_log.append({
+                    "Task ID": task_id,
+                    "Status": "⚡ Cached",
+                    "Question": question_text[:50] + "...",
+                    "Answer": answer
+                })
+                payload_cache.append({"task_id": task_id, "submitted_answer": answer})
+                yield f"⚡ Task {i+1} Cached.", pd.DataFrame(results_log)
+                continue
         
-        raw_content = result["messages"][-1].content
+        # --- REAL GENERATION (Slow) ---
+        final_answer = "NOT KNOWN"
+        success = False
         
-        # --- FASE DI PULIZIA ---
-        clean_content = self._clean_answer(raw_content)
+        for attempt in range(MAX_RETRIES):
+            try:
+                status_msg = f"⏳ Task {i+1}/{total} (Try {attempt+1})..."
+                yield status_msg, pd.DataFrame(results_log)
+                
+                answer = agent(question_text)
+                
+                if is_valid_answer(answer):
+                    final_answer = answer
+                    success = True
+                    # Cooling down
+                    wait_time = REQUEST_DELAY
+                    yield f"✅ Success. Cooling down for {wait_time}s...", pd.DataFrame(results_log)
+                    time.sleep(wait_time) 
+                    break 
+                else:
+                    print(f"⚠️ Invalid answer: {answer}")
+                    yield f"⚠️ Invalid format. Retrying in {ERROR_DELAY}s...", pd.DataFrame(results_log)
+                    time.sleep(ERROR_DELAY)
+                    
+            except Exception as e:
+                print(f"❌ Error on attempt {attempt+1}: {e}")
+                yield f"❌ Error. Pausing {ERROR_DELAY}s...", pd.DataFrame(results_log)
+                time.sleep(ERROR_DELAY)
+
+        # Save Result
+        status_label = "✅ Generated" if success else "⚠️ Failed"
+        results_log.append({
+            "Task ID": task_id,
+            "Status": status_label,
+            "Question": question_text[:50] + "...",
+            "Answer": final_answer
+        })
         
-        # Debug print per vedere cosa stiamo tagliando (utile per te)
-        if raw_content != clean_content:
-            print(f"CLEANED OUTPUT:\nRaw: '{raw_content}'\nClean: '{clean_content}'")
-
+        payload_cache.append({
+            "task_id": task_id,
+            "submitted_answer": final_answer
+        })
+        
+        save_cache(payload_cache)
+        yield f"🏁 Logic Complete Task {i+1}", pd.DataFrame(results_log)
 
-def run_and_submit_all( profile: gr.OAuthProfile | None):
-    """
-    Fetches all questions, runs the GAIAgent 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
+    yield "🎉 Generation Complete! Ready to Submit.", pd.DataFrame(results_log)
 
-    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"
+# =============================================================================
+# 2. SUBMIT FUNCTION
+# =============================================================================
+def submit_to_leaderboard(profile: gr.OAuthProfile | None):
+    if not profile: return "⚠️ Please login first!"
+    
+    answers = load_cache_as_dict()
+    if not answers: return "⚠️ No answers found."
 
-    # 1. Instantiate Agent ( modify this part to create your agent)
-    try:
-        agent = GAIAgent()
-    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)
+    payload_list = [{"task_id": k, "submitted_answer": v} for k, v in answers.items()]
+    username = profile.username
+    space_id = os.getenv("SPACE_ID", "local-test")
     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)
+    
+    submission_data = {
+        "username": username,
+        "agent_code": agent_code,
+        "answers": payload_list
+    }
 
-    # 5. Submit
-    print(f"Submitting {len(answers_payload)} answers to: {submit_url}")
     try:
-        response = requests.post(submit_url, json=submission_data, timeout=60)
+        response = requests.post(f"{DEFAULT_API_URL}/submit", 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
+        res = response.json()
+        return f"✅ SUCCESS! Score: {res.get('score')}% ({res.get('correct_count')} correct)"
     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("# GAIAgent 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.
+        return f"❌ Submission Failed: {str(e)}"
 
-        ---
-        **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)
+# =============================================================================
+# 3. INTERFACE
+# =============================================================================
+with gr.Blocks(title="GAIA Agent Evaluation", theme=gr.themes.Soft()) as demo:
+    gr.Markdown("# 🤖 GAIA Agent Evaluation (Robust)")
+    gr.Markdown(f"**Settings:** Grace Time: {REQUEST_DELAY}s | Max Retries: {MAX_RETRIES}")
+    
+    with gr.Row(): login_btn = gr.LoginButton()
+    
+    with gr.Row():
+        btn_gen = gr.Button("1. Generate Answers (Safe Mode)", variant="primary")
+        btn_sub = gr.Button("2. Submit Results", variant="secondary")
 
-    run_button.click(
-        fn=run_and_submit_all,
-        outputs=[status_output, results_table]
-    )
+    status = gr.Textbox(label="Status", interactive=False)
+    table = gr.DataFrame(label="Results", wrap=True)
+    
+    btn_gen.click(fn=generate_answers, inputs=[login_btn], outputs=[status, table])
+    btn_sub.click(fn=submit_to_leaderboard, inputs=[login_btn], outputs=[status])
 
 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 GAIAgent Evaluation...")
-    demo.launch(debug=True, share=False)
+    demo.launch()

appOLD.py

@@ -0,0 +1,265 @@
+import os
+import re
+import gradio as gr
+import requests
+import inspect
+import pandas as pd
+
+from langchain_core.messages import HumanMessage
+from agent import build_graph
+
+# (Keep Constants as is)
+# --- Constants ---
+DEFAULT_API_URL = "https://agents-course-unit4-scoring.hf.space"
+
+# --- GAIAgent Definition ---
+# ----- THIS IS WERE YOU CAN BUILD WHAT YOU WANT ------
+class GAIAgent:
+    def __init__(self):
+        self.graph = build_graph(provider = "groq")
+
+    
+    def _clean_answer(self, text: str) -> str:
+        """
+        Pulisce la risposta per garantire l'Exact Match.
+        Strategia:
+        1. Se c'è qualcosa tra **doppi asterischi**, è probabilmente la risposta.
+        2. Se c'è 'FINAL ANSWER:', prendi quello che segue.
+        3. Pulisci markdown e punteggiatura.
+        """
+        if not text: 
+            return ""
+
+        # --- STRATEGIA 1: CATTURA GRASSETTO (Priorità Alta) ---
+        # Se l'LLM scrive "The answer is **128**", prendiamo solo "128".
+        bold_match = re.search(r"\*\*(.*?)\*\*", text)
+        if bold_match:
+            return bold_match.group(1).strip()
+        
+        # --- STRATEGIA 2: CATTURA PREFIX (Standard) ---
+        match = re.search(r"FINAL\s*ANSWER\s*:\s*(.*)", text, re.IGNORECASE | re.DOTALL)
+        if match: 
+            text = match.group(1).strip()
+
+        # --- PULIZIA GENERALE ---
+        # Rimuove markdown residuo
+        text = text.replace("**", "").replace("`", "").strip()
+        
+        # Prende solo la prima riga (spesso le spiegazioni sono sotto)
+        lines = [L.strip() for L in text.split('\n') if L.strip()]
+        if lines: text = lines[0]
+
+        # --- STRATEGIA 3: HEURISTICA "THE ANSWER IS..." ---
+        # Se dopo tutto ciò abbiamo ancora una frase lunga che finisce con il codice
+        # Esempio: "The number is 80GSFC21M0002."
+        # Proviamo a splittare su " is " o " are "
+        if len(text.split()) > 5: # Se è una frase lunga
+            split_match = re.search(r"\b(is|are|was|were)\b\s*(.*)$", text, re.IGNORECASE)
+            if split_match:
+                candidate = split_match.group(2).strip()
+                # Se quello che rimane è breve (la risposta), prendiamo quello
+                if len(candidate.split()) <= 3:
+                    text = candidate
+
+        # Rimuove punto finale
+        if text.endswith(".") and len(text) > 4: 
+            text = text[:-1]
+            
+        return text
+    
+
+    def __call__(self, question: str) -> str:
+        from langchain_core.messages import HumanMessage
+        messages = [HumanMessage(content=question)]
+        
+        # Eseguiamo il grafo con un limite di ricorsione alto
+        result = self.graph.invoke(
+            {"messages": messages},
+            config={"recursion_limit": 100} 
+        )
+        
+        raw_content = result["messages"][-1].content
+        
+        # --- FASE DI PULIZIA ---
+        clean_content = self._clean_answer(raw_content)
+        
+        # Debug print per vedere cosa stiamo tagliando (utile per te)
+        if raw_content != clean_content:
+            print(f"CLEANED OUTPUT:\nRaw: '{raw_content}'\nClean: '{clean_content}'")
+
+
+def run_and_submit_all( profile: gr.OAuthProfile | None):
+    """
+    Fetches all questions, runs the GAIAgent 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 = GAIAgent()
+    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("# GAIAgent 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 GAIAgent Evaluation...")
+    demo.launch(debug=True, share=False)