updated agent with more tools

app.py

@@ -4,7 +4,7 @@ import requests
 import inspect
 import pandas as pd
 from dotenv import load_dotenv
-from huggingface_hub import InferenceClient
+from smolagents import CodeAgent, DuckDuckGoSearchTool, FinalAnswerTool, InferenceClientModel, Tool, tool, VisitWebpageTool
 
 # Load environment variables
 load_dotenv()
@@ -12,9 +12,92 @@ load_dotenv()
 # --- Constants ---
 DEFAULT_API_URL = "https://agents-course-unit4-scoring.hf.space"
 
+# --- Custom Tools for GAIA Dataset ---
+
+@tool
+def calculate_math(expression: str) -> str:
+    """
+    Calculates mathematical expressions safely.
+    Args:
+        expression: Mathematical expression to evaluate (e.g., "2 + 2", "sqrt(16)")
+    """
+    try:
+        import math
+        import re
+        
+        # Replace common math functions
+        expression = expression.replace("sqrt", "math.sqrt")
+        expression = expression.replace("log", "math.log")
+        expression = expression.replace("sin", "math.sin")
+        expression = expression.replace("cos", "math.cos")
+        expression = expression.replace("tan", "math.tan")
+        expression = expression.replace("pi", "math.pi")
+        expression = expression.replace("e", "math.e")
+        
+        # Safe evaluation
+        allowed_names = {
+            k: v for k, v in math.__dict__.items() if not k.startswith("__")
+        }
+        allowed_names.update({"abs": abs, "round": round, "min": min, "max": max})
+        
+        result = eval(expression, {"__builtins__": {}}, allowed_names)
+        return str(result)
+    except Exception as e:
+        return f"Error calculating: {str(e)}"
+
+@tool
+def analyze_data(data_description: str) -> str:
+    """
+    Analyzes data patterns, statistics, or trends described in text.
+    Args:
+        data_description: Description of data to analyze
+    """
+    # This is a simplified analysis tool
+    # In a real scenario, this could connect to data analysis libraries
+    return f"Data analysis for: {data_description}. Please provide specific data or use web search for current statistics."
+
+@tool
+def fact_checker(claim: str) -> str:
+    """
+    Helps verify factual claims by suggesting verification approaches.
+    Args:
+        claim: The factual claim to verify
+    """
+    return f"To verify '{claim}', I recommend using web search for recent, authoritative sources. Cross-reference multiple reliable sources."
+
+class AdvancedReasoningTool(Tool):
+    name = "advanced_reasoning"
+    description = """
+    This tool helps break down complex multi-step reasoning problems.
+    It provides structured thinking for complex questions."""
+    
+    inputs = {
+        "problem": {
+            "type": "string",
+            "description": "A complex problem that requires step-by-step reasoning",
+        },
+        "problem_type": {
+            "type": "string", 
+            "description": "Type of problem (e.g., 'logical', 'mathematical', 'analytical', 'research')",
+        }
+    }
+    
+    output_type = "string"
+
+    def forward(self, problem: str, problem_type: str = "general"):
+        reasoning_frameworks = {
+            "logical": "1. Identify premises\n2. Apply logical rules\n3. Check for contradictions\n4. Draw conclusions",
+            "mathematical": "1. Understand what's being asked\n2. Identify known values\n3. Choose appropriate formulas\n4. Calculate step-by-step\n5. Verify the answer",
+            "analytical": "1. Break down into components\n2. Analyze each part\n3. Look for patterns/relationships\n4. Synthesize findings",
+            "research": "1. Define research question\n2. Identify reliable sources\n3. Gather information\n4. Cross-reference facts\n5. Form conclusion"
+        }
+        
+        framework = reasoning_frameworks.get(problem_type.lower(), reasoning_frameworks["analytical"])
+        return f"Problem: {problem}\n\nSuggested approach ({problem_type}):\n{framework}"
+
 class BasicAgent:
     def __init__(self):
-        print("BasicAgent initialized - using HF Inference API.")
+        print("🤖 BasicAgent initialized with smolagents framework.")
         
         # Get HF token from environment
         hf_token = os.getenv("HUGGINGFACE_HUB_TOKEN") or os.getenv("HF_TOKEN")
@@ -23,64 +106,80 @@ class BasicAgent:
             raise ValueError("❌ No HF token found. Please set HF_TOKEN or HUGGINGFACE_HUB_TOKEN in your .env file\n"
                            "You can get a token from: https://huggingface.co/settings/tokens")
         
-        # Initialize the Inference Client
         try:
-            self.client = InferenceClient(
-                provider="hf-inference",
-                api_key=hf_token,
+            # Initialize model with HF token
+            model = InferenceClientModel(
+                model_id="HuggingFaceTB/SmolLM3-3B",
+                token=hf_token
+            )
+            
+            # Create agent with comprehensive tools
+            self.agent = CodeAgent(
+                tools=[
+                    DuckDuckGoSearchTool(),
+                    VisitWebpageTool(),
+                    calculate_math,
+                    analyze_data,
+                    fact_checker,
+                    AdvancedReasoningTool(),
+                    FinalAnswerTool()
+                ],
+                model=model,
+                max_steps=15,  # Increased for complex GAIA questions
+                verbosity_level=2
             )
-            print("✅ HF Inference Client initialized successfully")
+            print("✅ SmolAgent initialized successfully with all tools")
+            
         except Exception as e:
-            print(f"❌ Error initializing HF Inference Client: {e}")
+            print(f"❌ Error initializing SmolAgent: {e}")
             raise e
     
     def __call__(self, question: str) -> str:
-        print(f"🤖 Agent received question: {question[:100]}...")
+        print(f"🤖 SmolAgent received question: {question[:100]}...")
         
         try:
-            print("🔄 Making API call to HF Inference...")
+            print("🔄 Running SmolAgent with tools...")
             
-            # Create a more detailed prompt for better responses
-            system_prompt = "You are a helpful AI assistant. Provide clear, accurate, and concise answers to questions."
+            # Add context to help the agent understand it should provide a final answer
+            enhanced_question = f"""
+            Please answer the following question thoroughly and accurately. Use the available tools to search for information, visit websites, perform calculations, or analyze data as needed.
+
+            Question: {question}
+
+            Please provide a clear, specific final answer at the end.
+            """
             
-            completion = self.client.chat.completions.create(
-                model="HuggingFaceTB/SmolLM3-3B",
-                messages=[
-                    {
-                        "role": "system",
-                        "content": system_prompt
-                    },
-                    {
-                        "role": "user",
-                        "content": question
-                    }
-                ],
-                max_tokens=512,
-                temperature=0.7,
-            )
+            result = self.agent.run(enhanced_question)
+            
+            print("✅ SmolAgent completed successfully!")
             
-            print("✅ API call successful!")
-            answer = completion.choices[0].message.content.strip()
-            print(f"📝 Agent returning answer: {answer[:200]}...")
+            # Extract the final answer if it's wrapped in agent output
+            if hasattr(result, 'content'):
+                answer = result.content
+            elif isinstance(result, dict) and 'output' in result:
+                answer = result['output']
+            else:
+                answer = str(result)
             
-            # Ensure we're not returning empty or None answers
+            print(f"📝 SmolAgent returning answer: {answer[:200]}...")
+            
+            # Ensure we have a meaningful answer
             if not answer or answer.lower().strip() == "":
                 return "I apologize, but I couldn't generate a proper response to your question."
             
             return answer
             
         except Exception as e:
-            error_msg = f"❌ HF API Error: {str(e)}"
+            error_msg = f"❌ SmolAgent Error: {str(e)}"
             print(error_msg)
             print(f"📋 Full error details: {repr(e)}")
             
-            # Return a more informative error message
             return f"Sorry, I encountered an error while processing your question: {str(e)}"
     
     def test_connection(self):
-        """Test if the HF API connection is working"""
+        """Test if the agent is working properly"""
         try:
-            test_response = self("What is 2+2?")
+            test_response = self("What is the capital of France?")
             print(f"🧪 Test response: {test_response}")
             return True, test_response
         except Exception as e:
@@ -116,7 +215,7 @@ def run_and_submit_all(profile: gr.OAuthProfile | None):
 
     # 1. Instantiate Agent
     try:
-        print("🚀 Initializing BasicAgent...")
+        print("🚀 Initializing SmolAgent...")
         agent = BasicAgent()
         
         # Test the agent before proceeding
@@ -156,7 +255,7 @@ def run_and_submit_all(profile: gr.OAuthProfile | None):
     # 3. Run your Agent
     results_log = []
     answers_payload = []
-    print(f"Running agent on {len(questions_data)} questions...")
+    print(f"Running SmolAgent on {len(questions_data)} questions...")
     
     for i, item in enumerate(questions_data):
         task_id = item.get("task_id")
@@ -165,18 +264,19 @@ def run_and_submit_all(profile: gr.OAuthProfile | None):
             print(f"Skipping item with missing task_id or question: {item}")
             continue
         
-        print(f"Processing question {i+1}/{len(questions_data)}: {task_id}")
+        print(f"🔄 Processing question {i+1}/{len(questions_data)}: {task_id}")
         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[:100] + "..." if len(question_text) > 100 else question_text, 
-                "Submitted Answer": submitted_answer[:200] + "..." if len(submitted_answer) > 200 else submitted_answer
+                "Submitted Answer": submitted_answer[:300] + "..." if len(submitted_answer) > 300 else submitted_answer
             })
+            print(f"✅ Completed question {i+1}")
         except Exception as e:
             error_msg = f"AGENT ERROR: {e}"
-            print(f"Error running agent on task {task_id}: {e}")
+            print(f"❌ Error running agent on task {task_id}: {e}")
             answers_payload.append({"task_id": task_id, "submitted_answer": error_msg})
             results_log.append({
                 "Task ID": task_id, 
@@ -194,7 +294,7 @@ def run_and_submit_all(profile: gr.OAuthProfile | None):
         "agent_code": agent_code, 
         "answers": answers_payload
     }
-    status_update = f"Agent finished. Submitting {len(answers_payload)} answers for user '{username}'..."
+    status_update = f"SmolAgent finished. Submitting {len(answers_payload)} answers for user '{username}'..."
     print(status_update)
 
     # 5. Submit
@@ -243,19 +343,24 @@ def run_and_submit_all(profile: gr.OAuthProfile | None):
 
 # --- Build Gradio Interface using Blocks ---
 with gr.Blocks() as demo:
-    gr.Markdown("# Basic Agent Evaluation Runner")
+    gr.Markdown("# SmolAgent Evaluation Runner for GAIA")
     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.
+        1. This agent uses smolagents framework with multiple tools:
+           - 🔍 **DuckDuckGoSearchTool**: Web search capabilities
+           - 🌐 **VisitWebpageTool**: Can visit and read web pages
+           - 🧮 **Math Calculator**: Handles mathematical calculations
+           - 📊 **Data Analysis**: Basic data analysis capabilities
+           - ✅ **Fact Checker**: Helps verify claims
+           - 🧠 **Advanced Reasoning**: Structured problem-solving
+        
+        2. Log in to your Hugging Face account using the button below.
+        3. Click 'Run Evaluation & Submit All Answers' to start the evaluation.
 
         ---
-        **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.
+        **Note:** This agent is designed for the GAIA dataset and can handle complex, multi-step reasoning tasks.
         """
     )
 
@@ -272,7 +377,7 @@ with gr.Blocks() as demo:
     )
 
 if __name__ == "__main__":
-    print("\n" + "-"*30 + " App Starting " + "-"*30)
+    print("\n" + "-"*30 + " SmolAgent 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
@@ -290,7 +395,7 @@ if __name__ == "__main__":
     else:
         print("ℹ️  SPACE_ID environment variable not found (running locally?). Repo URL cannot be determined.")
 
-    print("-"*(60 + len(" App Starting ")) + "\n")
+    print("-"*(60 + len(" SmolAgent Starting ")) + "\n")
 
-    print("Launching Gradio Interface for Basic Agent Evaluation...")
+    print("Launching Gradio Interface for SmolAgent GAIA Evaluation...")
     demo.launch(debug=True, share=False)