Update app.py

app.py

@@ -3,7 +3,7 @@ import re
 import requests
 import gradio as gr
 import pandas as pd
-from smolagents import CodeAgent, DuckDuckGoSearchTool, InferenceClientModel, tool
+from smolagents import CodeAgent, tool, LiteLLMModel
 
 # --- Constants ---
 DEFAULT_API_URL = "https://agents-course-unit4-scoring.hf.space"
@@ -25,16 +25,15 @@ def calculator(expression: str) -> str:
     """
     import math
     try:
-        # Clean the expression
         expression = expression.strip()
         
-        # Create safe math context
         safe_dict = {
             "abs": abs, "round": round, "min": min, "max": max,
             "sum": sum, "pow": pow, "len": len,
             "sqrt": math.sqrt, "sin": math.sin, "cos": math.cos,
             "tan": math.tan, "log": math.log, "log10": math.log10,
             "pi": math.pi, "e": math.e, "floor": math.floor, "ceil": math.ceil,
+            "factorial": math.factorial,
         }
         
         result = eval(expression, {"__builtins__": {}}, safe_dict)
@@ -87,7 +86,6 @@ def visit_webpage(url: str) -> str:
         The text content of the webpage (truncated if too long)
     """
     try:
-        import requests
         from bs4 import BeautifulSoup
         
         headers = {
@@ -99,15 +97,13 @@ def visit_webpage(url: str) -> str:
         
         soup = BeautifulSoup(response.text, 'html.parser')
         
-        # Remove script and style elements
         for element in soup(['script', 'style', 'nav', 'footer', 'header']):
             element.decompose()
         
         text = soup.get_text(separator='\n', strip=True)
         
-        # Truncate if too long
-        if len(text) > 8000:
-            text = text[:8000] + "\n...[truncated]"
+        if len(text) > 10000:
+            text = text[:10000] + "\n...[truncated]"
         
         return text if text else "Could not extract text from webpage."
     except Exception as e:
@@ -126,9 +122,6 @@ def wikipedia_search(query: str) -> str:
         Wikipedia article summary and key information
     """
     try:
-        import requests
-        
-        # Search Wikipedia API
         search_url = "https://en.wikipedia.org/w/api.php"
         search_params = {
             "action": "query",
@@ -144,14 +137,13 @@ def wikipedia_search(query: str) -> str:
         if not data.get("query", {}).get("search"):
             return f"No Wikipedia articles found for: {query}"
         
-        # Get the first result's page content
         title = data["query"]["search"][0]["title"]
         
         content_params = {
             "action": "query",
             "titles": title,
             "prop": "extracts",
-            "exintro": True,
+            "exintro": False,
             "explaintext": True,
             "format": "json"
         }
@@ -162,8 +154,8 @@ def wikipedia_search(query: str) -> str:
         pages = data.get("query", {}).get("pages", {})
         for page_id, page_data in pages.items():
             extract = page_data.get("extract", "No content available")
-            if len(extract) > 3000:
-                extract = extract[:3000] + "...[truncated]"
+            if len(extract) > 5000:
+                extract = extract[:5000] + "...[truncated]"
             return f"Wikipedia: {title}\n\n{extract}"
         
         return "Could not retrieve Wikipedia content."
@@ -171,54 +163,11 @@ def wikipedia_search(query: str) -> str:
         return f"Wikipedia error: {str(e)}"
 
 
-@tool
-def read_file_from_url(url: str) -> str:
-    """
-    Downloads and reads content from a file URL (txt, csv, json, etc).
-    
-    Args:
-        url: The URL of the file to download and read
-    
-    Returns:
-        The content of the file
-    """
-    try:
-        import requests
-        
-        headers = {"User-Agent": "Mozilla/5.0"}
-        response = requests.get(url, headers=headers, timeout=30)
-        response.raise_for_status()
-        
-        content_type = response.headers.get('content-type', '').lower()
-        
-        # Handle different file types
-        if 'json' in content_type or url.endswith('.json'):
-            import json
-            data = response.json()
-            return json.dumps(data, indent=2)[:5000]
-        
-        elif 'csv' in content_type or url.endswith('.csv'):
-            lines = response.text.split('\n')[:50]  # First 50 lines
-            return '\n'.join(lines)
-        
-        elif url.endswith('.xlsx') or url.endswith('.xls'):
-            return "Excel file detected. Cannot read directly - need pandas with openpyxl."
-        
-        else:
-            # Plain text
-            text = response.text
-            if len(text) > 5000:
-                text = text[:5000] + "\n...[truncated]"
-            return text
-            
-    except Exception as e:
-        return f"Error reading file: {str(e)}"
-
-
 @tool
 def get_gaia_file(task_id: str) -> str:
     """
     Downloads a file associated with a GAIA task from the API.
+    Use this tool when the question mentions a file or attachment.
     
     Args:
         task_id: The task ID to get the file for
@@ -227,8 +176,6 @@ def get_gaia_file(task_id: str) -> str:
         Information about the file or its content if text-based
     """
     try:
-        import requests
-        
         api_url = f"https://agents-course-unit4-scoring.hf.space/files/{task_id}"
         response = requests.get(api_url, timeout=30)
         
@@ -240,41 +187,82 @@ def get_gaia_file(task_id: str) -> str:
         content_type = response.headers.get('content-type', '').lower()
         content_disp = response.headers.get('content-disposition', '')
         
-        # Try to get filename
         filename = "unknown"
         if 'filename=' in content_disp:
             filename = content_disp.split('filename=')[-1].strip('"\'')
         
-        # Handle based on content type
-        if 'text' in content_type or 'json' in content_type:
+        # Handle text files
+        if 'text' in content_type or filename.endswith(('.txt', '.csv', '.json', '.md')):
             content = response.text
-            if len(content) > 5000:
-                content = content[:5000] + "\n...[truncated]"
-            return f"File: {filename}\nContent:\n{content}"
-        
-        elif 'image' in content_type:
-            return f"File: {filename}\nType: Image file ({content_type})\nNote: This is an image file. I cannot view images directly."
-        
-        elif 'audio' in content_type:
-            return f"File: {filename}\nType: Audio file ({content_type})\nNote: This is an audio file. I cannot process audio directly."
+            if len(content) > 8000:
+                content = content[:8000] + "\n...[truncated]"
+            return f"File: {filename}\n\nContent:\n{content}"
         
+        # Handle Python files
+        elif filename.endswith('.py'):
+            content = response.text
+            if len(content) > 8000:
+                content = content[:8000] + "\n...[truncated]"
+            return f"Python File: {filename}\n\nCode:\n{content}"
+        
+        # Handle Excel files
+        elif filename.endswith(('.xlsx', '.xls')):
+            try:
+                import pandas as pd
+                from io import BytesIO
+                df = pd.read_excel(BytesIO(response.content))
+                return f"Excel File: {filename}\n\nData:\n{df.to_string()}"
+            except:
+                return f"Excel File: {filename}\nNote: Could not parse Excel file."
+        
+        # Handle images
+        elif 'image' in content_type or filename.endswith(('.png', '.jpg', '.jpeg', '.gif')):
+            return f"File: {filename}\nType: Image ({content_type})\nNote: This is an image file. I cannot view images directly, but I can tell you it exists."
+        
+        # Handle audio
+        elif 'audio' in content_type or filename.endswith(('.mp3', '.wav', '.m4a')):
+            return f"File: {filename}\nType: Audio ({content_type})\nNote: This is an audio file. I cannot process audio directly."
+        
+        # Handle PDF
         elif 'pdf' in content_type or filename.endswith('.pdf'):
             return f"File: {filename}\nType: PDF document\nNote: This is a PDF file. I cannot read PDFs directly."
         
-        elif 'spreadsheet' in content_type or filename.endswith(('.xlsx', '.xls')):
-            return f"File: {filename}\nType: Excel spreadsheet\nNote: This is an Excel file. I cannot read Excel directly."
-        
         else:
-            return f"File: {filename}\nType: {content_type}\nSize: {len(response.content)} bytes\nNote: Binary file, cannot display content."
+            return f"File: {filename}\nType: {content_type}\nSize: {len(response.content)} bytes"
             
     except Exception as e:
         return f"Error getting file: {str(e)}"
 
 
+@tool
+def read_file_content(url: str) -> str:
+    """
+    Downloads and reads content from a file URL.
+    
+    Args:
+        url: The URL of the file to download and read
+    
+    Returns:
+        The content of the file
+    """
+    try:
+        headers = {"User-Agent": "Mozilla/5.0"}
+        response = requests.get(url, headers=headers, timeout=30)
+        response.raise_for_status()
+        
+        content = response.text
+        if len(content) > 8000:
+            content = content[:8000] + "\n...[truncated]"
+        return content
+            
+    except Exception as e:
+        return f"Error reading file: {str(e)}"
+
+
 @tool
 def reverse_text(text: str) -> str:
     """
-    Reverses the given text string.
+    Reverses the given text string character by character.
     
     Args:
         text: The text to reverse
@@ -301,12 +289,12 @@ def count_items(text: str, item_type: str = "words") -> str:
     
     if item_type == "words":
         count = len(text.split())
-    elif item_type == "characters" or item_type == "chars":
+    elif item_type in ["characters", "chars", "char"]:
         count = len(text)
     elif item_type == "lines":
         count = len(text.split('\n'))
     elif item_type == "sentences":
-        count = len(re.split(r'[.!?]+', text))
+        count = len(re.split(r'[.!?]+', text.strip()))
     else:
         return f"Unknown item type: {item_type}. Use: words, characters, lines, or sentences."
     
@@ -333,7 +321,7 @@ def extract_numbers(text: str) -> str:
 @tool
 def sort_list(items: str, order: str = "ascending") -> str:
     """
-    Sorts a comma-separated list of items.
+    Sorts a comma-separated list of items alphabetically or numerically.
     
     Args:
         items: Comma-separated items to sort (e.g., "banana, apple, cherry")
@@ -344,13 +332,11 @@ def sort_list(items: str, order: str = "ascending") -> str:
     """
     item_list = [item.strip() for item in items.split(',')]
     
-    # Try to sort as numbers if all items are numeric
     try:
         numeric_list = [float(item) for item in item_list]
         sorted_list = sorted(numeric_list, reverse=(order.lower() == "descending"))
-        return ', '.join(str(int(x) if x.is_integer() else x) for x in sorted_list)
+        return ', '.join(str(int(x) if x == int(x) else x) for x in sorted_list)
     except ValueError:
-        # Sort as strings
         sorted_list = sorted(item_list, reverse=(order.lower() == "descending"))
         return ', '.join(sorted_list)
 
@@ -358,7 +344,7 @@ def sort_list(items: str, order: str = "ascending") -> str:
 @tool
 def convert_units(value: float, from_unit: str, to_unit: str) -> str:
     """
-    Converts between common units.
+    Converts between common units of measurement.
     
     Args:
         value: The numeric value to convert
@@ -366,115 +352,138 @@ def convert_units(value: float, from_unit: str, to_unit: str) -> str:
         to_unit: The target unit
     
     Returns:
-        The converted value
+        The converted value with units
     """
     conversions = {
-        # Length
         ("km", "miles"): lambda x: x * 0.621371,
         ("miles", "km"): lambda x: x * 1.60934,
         ("m", "feet"): lambda x: x * 3.28084,
         ("feet", "m"): lambda x: x * 0.3048,
         ("cm", "inches"): lambda x: x * 0.393701,
         ("inches", "cm"): lambda x: x * 2.54,
-        # Temperature
         ("celsius", "fahrenheit"): lambda x: (x * 9/5) + 32,
         ("fahrenheit", "celsius"): lambda x: (x - 32) * 5/9,
         ("celsius", "kelvin"): lambda x: x + 273.15,
         ("kelvin", "celsius"): lambda x: x - 273.15,
-        # Weight
         ("kg", "lbs"): lambda x: x * 2.20462,
         ("lbs", "kg"): lambda x: x * 0.453592,
         ("g", "oz"): lambda x: x * 0.035274,
         ("oz", "g"): lambda x: x * 28.3495,
     }
     
-    key = (from_unit.lower(), to_unit.lower())
+    key = (from_unit.lower().strip(), to_unit.lower().strip())
     if key in conversions:
         result = conversions[key](value)
-        return f"{value} {from_unit} = {result:.4f} {to_unit}"
+        return f"{value} {from_unit} = {result:.6f} {to_unit}"
     else:
         return f"Conversion from {from_unit} to {to_unit} not supported."
 
 
+@tool
+def get_current_time() -> str:
+    """
+    Gets the current date and time in UTC.
+    
+    Returns:
+        The current date and time
+    """
+    from datetime import datetime
+    now = datetime.utcnow()
+    return f"Current UTC date/time: {now.strftime('%Y-%m-%d %H:%M:%S')}"
+
+
 # ============================================
-# BASIC AGENT CLASS - USING SMOLAGENTS
+# BASIC AGENT CLASS - USING GROQ
 # ============================================
 
 class BasicAgent:
     def __init__(self):
-        print("Initializing BasicAgent with smolagents...")
+        print("Initializing BasicAgent with Groq + Llama 3.3 70B...")
         
-        # Use a larger model for better reasoning
-        self.model = InferenceClientModel(
-            model_id="Qwen/Qwen2.5-72B-Instruct",  # Bigger model
-            token=os.environ.get("HF_TOKEN"),
+        # Use Groq with Llama 3.3 70B - fast and smart!
+        self.model = LiteLLMModel(
+            model_id="groq/llama-3.3-70b-versatile",
+            api_key=os.environ.get("GROQ_API_KEY"),
         )
         
-        # Create the agent with comprehensive tools
+        # Create the agent with all tools
         self.agent = CodeAgent(
             model=self.model,
             tools=[
-                web_search,          # Web search
-                visit_webpage,       # Visit and read webpages
-                wikipedia_search,    # Wikipedia lookup
-                calculator,          # Math calculations
-                read_file_from_url,  # Read files from URLs
-                get_gaia_file,       # Get GAIA task files
-                reverse_text,        # Reverse strings
-                count_items,         # Count words/chars/lines
-                extract_numbers,     # Extract numbers from text
-                sort_list,           # Sort lists
-                convert_units,       # Unit conversions
+                web_search,
+                visit_webpage,
+                wikipedia_search,
+                calculator,
+                get_gaia_file,
+                read_file_content,
+                reverse_text,
+                count_items,
+                extract_numbers,
+                sort_list,
+                convert_units,
+                get_current_time,
             ],
-            max_steps=12,  # More steps for complex reasoning
+            max_steps=15,
             verbosity_level=1,
         )
         
-        print("BasicAgent initialized successfully!")
+        print("BasicAgent initialized successfully with Groq!")
     
     def __call__(self, question: str, task_id: str = None) -> str:
-        print(f"Agent received question: {question[:100]}...")
+        print(f"Agent processing: {question[:100]}...")
         
         try:
-            # Build enhanced prompt
-            file_note = ""
+            # Build the prompt with clear instructions
+            file_instruction = ""
             if task_id:
-                file_note = f"\n\nNote: This task may have an associated file. Use get_gaia_file('{task_id}') to check."
-            
-            enhanced_prompt = f"""You are an AI assistant solving questions for the GAIA benchmark.
-
-IMPORTANT INSTRUCTIONS:
-1. Think step by step before answering
-2. Use tools when needed to find information
-3. If the question mentions a file, use get_gaia_file() with the task_id
-4. For web lookups, use web_search() first, then visit_webpage() if needed
-5. For factual info, try wikipedia_search()
-6. Double-check calculations with the calculator tool
-7. Give ONLY the final answer - no explanations
-8. Keep answers concise and exact (the grading uses exact match)
-{file_note}
+                file_instruction = f"""
+IMPORTANT: This question may have an associated file. 
+To check for and read the file, use: get_gaia_file("{task_id}")
+Always check for a file first if the question mentions any attachment, file, document, image, or data."""
+
+            enhanced_prompt = f"""You are solving a GAIA benchmark question. Follow these rules:
+
+1. THINK step by step before answering
+2. USE TOOLS when you need information:
+   - web_search() for current info or facts
+   - wikipedia_search() for encyclopedic knowledge  
+   - visit_webpage() to read full webpage content
+   - calculator() for any math
+   - get_gaia_file("{task_id}") if there's an attached file
+3. VERIFY your answer before submitting
+4. Give ONLY the final answer - no explanation
+5. Be PRECISE - answers are graded by exact match
+{file_instruction}
 
 Question: {question}
 
-Think through this step-by-step, then provide your final answer."""
+Solve this step-by-step, then give your final answer."""
 
             # Run the agent
             answer = self.agent.run(enhanced_prompt)
             
-            # Clean up the answer
+            # Clean up answer
             answer = str(answer).strip()
             
-            # Remove common prefixes that might hurt exact matching
-            prefixes_to_remove = [
-                "The answer is: ", "The answer is ", 
-                "Answer: ", "Final answer: ",
+            # Remove common prefixes
+            prefixes = [
+                "The answer is: ", "The answer is ",
+                "Answer: ", "Final answer: ", "Final Answer: ",
                 "The final answer is: ", "The final answer is ",
+                "FINAL ANSWER: ", "FINAL ANSWER ",
             ]
-            for prefix in prefixes_to_remove:
-                if answer.lower().startswith(prefix.lower()):
+            for prefix in prefixes:
+                if answer.startswith(prefix):
+                    answer = answer[len(prefix):].strip()
+                elif answer.lower().startswith(prefix.lower()):
                     answer = answer[len(prefix):].strip()
             
-            print(f"Agent answer: {answer[:200]}...")
+            # Remove quotes if wrapped
+            if (answer.startswith('"') and answer.endswith('"')) or \
+               (answer.startswith("'") and answer.endswith("'")):
+                answer = answer[1:-1]
+            
+            print(f"Final answer: {answer[:200]}")
             return answer
             
         except Exception as e:
@@ -512,7 +521,7 @@ def run_and_submit_all(profile: gr.OAuthProfile | None):
         return f"Error initializing agent: {e}", None
     
     agent_code = f"https://huggingface.co/spaces/{space_id}/tree/main"
-    print(agent_code)
+    print(f"Agent code URL: {agent_code}")
 
     # 2. Fetch Questions
     print(f"Fetching questions from: {questions_url}")
@@ -521,92 +530,88 @@ def run_and_submit_all(profile: gr.OAuthProfile | None):
         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
+            return "Fetched questions list is empty.", None
         print(f"Fetched {len(questions_data)} questions.")
-    except requests.exceptions.RequestException as e:
+    except Exception as e:
         print(f"Error fetching questions: {e}")
         return f"Error fetching 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
+    # 3. Run Agent on all questions
     results_log = []
     answers_payload = []
+    print(f"\n{'='*60}")
     print(f"Running agent on {len(questions_data)} questions...")
+    print(f"{'='*60}\n")
     
     for i, item in enumerate(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
         
-        print(f"\n--- Question {i+1}/{len(questions_data)} ---")
-        print(f"Task ID: {task_id}")
-        print(f"Question: {question_text[:100]}...")
+        print(f"\n[{i+1}/{len(questions_data)}] Task: {task_id}")
+        print(f"Question: {question_text[:150]}{'...' if len(question_text) > 150 else ''}")
         
         try:
-            # Pass task_id so agent can fetch associated files
             submitted_answer = agent(question_text, task_id=task_id)
             answers_payload.append({"task_id": task_id, "submitted_answer": submitted_answer})
             results_log.append({
-                "Task ID": task_id, 
+                "Task ID": task_id,
                 "Question": question_text[:100] + "..." if len(question_text) > 100 else question_text,
-                "Submitted Answer": submitted_answer
+                "Answer": submitted_answer[:200] if len(submitted_answer) > 200 else submitted_answer
             })
-            print(f"Answer: {submitted_answer}")
+            print(f"✓ Answer: {submitted_answer[:100]}")
         except Exception as e:
-            print(f"Error running agent on task {task_id}: {e}")
+            print(f"✗ Error: {e}")
             results_log.append({
-                "Task ID": task_id, 
+                "Task ID": task_id,
                 "Question": question_text[:100] + "...",
-                "Submitted Answer": f"AGENT ERROR: {e}"
+                "Answer": f"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)
+        return "Agent did not produce any answers.", 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)
+    # 4. Submit answers
+    submission_data = {
+        "username": username.strip(),
+        "agent_code": agent_code,
+        "answers": answers_payload
+    }
+    
+    print(f"\n{'='*60}")
+    print(f"Submitting {len(answers_payload)} answers...")
+    print(f"{'='*60}\n")
 
-    # 5. Submit
-    print(f"Submitting {len(answers_payload)} answers to: {submit_url}")
     try:
-        response = requests.post(submit_url, json=submission_data, timeout=120)  # Longer timeout
+        response = requests.post(submit_url, json=submission_data, timeout=120)
         response.raise_for_status()
         result_data = response.json()
+        
+        score = result_data.get('score', 'N/A')
+        correct = result_data.get('correct_count', '?')
+        total = result_data.get('total_attempted', '?')
+        
         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.')}"
+            f"✅ Submission Successful!\n\n"
+            f"👤 User: {result_data.get('username')}\n"
+            f"🎯 Score: {score}% ({correct}/{total} correct)\n\n"
+            f"📝 {result_data.get('message', '')}"
         )
-        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:
-            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
+        
+        if float(score) >= 30:
+            final_status += "\n\n🎉 CONGRATULATIONS! You passed the 30% threshold!"
+        else:
+            final_status += f"\n\n📈 Need {30 - float(score)}% more to reach 30% passing score."
+        
+        print(final_status)
+        return final_status, pd.DataFrame(results_log)
+        
     except Exception as e:
-        status_message = f"An unexpected error occurred during submission: {e}"
+        status_message = f"Submission Failed: {e}"
         print(status_message)
-        results_df = pd.DataFrame(results_log)
-        return status_message, results_df
+        return status_message, pd.DataFrame(results_log)
 
 
 # ============================================
@@ -619,37 +624,31 @@ with gr.Blocks() as demo:
         """
         **Unit 4 Final Project - HuggingFace AI Agents Course**
         
-        This agent uses **smolagents** with **Qwen2.5-72B-Instruct** and the following tools:
-        
-        | Tool | Description |
-        |------|-------------|
-        | 🔍 **Web Search** | Search the internet via DuckDuckGo |
-        | 🌐 **Visit Webpage** | Extract text content from URLs |
-        | 📚 **Wikipedia** | Search and read Wikipedia articles |
-        | 🧮 **Calculator** | Perform math calculations |
-        | 📁 **File Reader** | Read files from URLs |
-        | 📎 **GAIA File** | Get task-associated files |
-        | 🔄 **Reverse Text** | Reverse strings |
-        | 📊 **Count Items** | Count words/chars/lines |
-        | 🔢 **Extract Numbers** | Find numbers in text |
-        | 📋 **Sort List** | Sort comma-separated items |
-        | ⚖️ **Convert Units** | Convert between units |
+        This agent uses **Groq + Llama 3.3 70B** with the following tools:
+        
+        | Category | Tools |
+        |----------|-------|
+        | 🔍 **Search** | Web Search, Wikipedia, Visit Webpage |
+        | 🧮 **Math** | Calculator, Unit Converter |
+        | 📁 **Files** | GAIA File Reader, URL File Reader |
+        | 📝 **Text** | Reverse, Count Items, Extract Numbers, Sort List |
+        | 🕐 **Utility** | Current Time |
         
         ---
         **Instructions:**
-        1. Log in with your Hugging Face account
-        2. Click 'Run Evaluation & Submit All Answers'
-        3. Wait for the agent to process all questions (may take 10-15 minutes)
-        4. View your score!
+        1. Make sure `GROQ_API_KEY` is set in Space secrets
+        2. Log in with your Hugging Face account  
+        3. Click the button and wait (~10-15 mins)
+        4. You need **30%** to pass!
         """
     )
 
     gr.LoginButton()
+    
+    run_button = gr.Button("🚀 Run Evaluation & Submit All Answers", variant="primary", size="lg")
 
-    run_button = gr.Button("🚀 Run Evaluation & Submit All Answers", variant="primary")
-
-    status_output = gr.Textbox(label="Run Status / Submission Result", lines=6, interactive=False)
-    results_table = gr.DataFrame(label="Questions and Agent Answers", wrap=True)
+    status_output = gr.Textbox(label="Status", lines=8, interactive=False)
+    results_table = gr.DataFrame(label="Results", wrap=True)
 
     run_button.click(
         fn=run_and_submit_all,
@@ -658,17 +657,18 @@ with gr.Blocks() as demo:
 
 if __name__ == "__main__":
     print("\n" + "="*60)
-    print("GAIA Agent Evaluation Runner - Starting")
+    print("🎯 GAIA Agent - Powered by Groq + Llama 3.3 70B")
     print("="*60)
     
-    space_host = os.getenv("SPACE_HOST")
+    # Check for API key
+    if os.environ.get("GROQ_API_KEY"):
+        print("✅ GROQ_API_KEY found")
+    else:
+        print("⚠️  GROQ_API_KEY not found - add it to Space secrets!")
+    
     space_id = os.getenv("SPACE_ID")
-
-    if space_host:
-        print(f"✅ SPACE_HOST: {space_host}")
     if space_id:
-        print(f"✅ SPACE_ID: {space_id}")
-        print(f"   Repo: https://huggingface.co/spaces/{space_id}")
+        print(f"✅ Space: https://huggingface.co/spaces/{space_id}")
     
     print("="*60 + "\n")