Update app.py

app.py

@@ -3,177 +3,130 @@ import gradio as gr
 import requests
 import pandas as pd
 from smolagents import CodeAgent, HfApiModel, tool
-import inspect
 
 # --- Constants ---
 DEFAULT_API_URL = "https://agents-course-unit4-scoring.hf.space"
 
-# ====================== CUSTOM TOOLS ======================
-# Add any tools you want here. The more useful ones you add, the better your score.
-
-# ====================== CUSTOM TOOLS ======================
-
+# ====================== TOOLS ======================
 @tool
 def web_search(query: str) -> str:
-    """Perform a web search using DuckDuckGo and return the top results.
+    """Perform a web search and return top results.
     
     Args:
         query: The search query to look up on the web.
-    
-    Returns:
-        str: A string containing the title and snippet of the top search results.
     """
     try:
         from duckduckgo_search import DDGS
         with DDGS() as ddgs:
             results = list(ddgs.text(query, max_results=5))
             if not results:
-                return "No search results found."
-            formatted = []
-            for r in results:
-                formatted.append(f"Title: {r.get('title', 'N/A')}\nSnippet: {r.get('body', r.get('snippet', 'N/A'))}")
-            return "\n\n".join(formatted)
+                return "No results found."
+            return "\n\n".join([f"Title: {r.get('title', '')}\nSnippet: {r.get('body') or r.get('snippet', '')}" for r in results])
     except Exception as e:
-        return f"Web search failed: {str(e)}"
+        return f"Search failed: {str(e)}"
 
 
 @tool
 def calculate(expression: str) -> str:
-    """Evaluate a simple mathematical expression.
+    """Evaluate a mathematical expression safely.
     
     Args:
-        expression: The math expression to calculate (e.g. '2 + 2 * 3', 'sin(3.14)', etc.)
-    
-    Returns:
-        str: The result of the calculation as a string.
+        expression: The math expression to calculate (e.g. "2 + 2 * 3")
     """
     try:
         import math
-        allowed_names = {"math": math}
-        return str(eval(expression, {"__builtins__": {}}, allowed_names))
+        return str(eval(expression, {"__builtins__": {}}, {"math": math}))
     except Exception as e:
         return f"Calculation error: {str(e)}"
 
 
-# Add more tools as needed (file handling, image description, code execution, etc.)
-# GAIA often requires: search, math, file reading, reasoning over tables/images, etc.
-
-# ====================== AGENT DEFINITION ======================
+# ====================== BASIC AGENT ======================
 class BasicAgent:
     def __init__(self):
-        print("🚀 Initializing Smolagents Agent for GAIA benchmark...")
-        
-        # Choose your model
-        # Option 1: Free HF Inference (good enough for many questions)
+        print("🚀 Initializing Smolagents Agent...")
+        # Using a strong free model on HF Inference
         self.model = HfApiModel(model_id="Qwen/Qwen2.5-Coder-32B-Instruct")
         
-        # Option 2: Stronger (if you have API key) → uncomment and set env var
-        # self.model = HfApiModel(
-        #     model_id="gpt-4o-mini",  # or "Qwen/Qwen2.5-72B-Instruct"
-        #     provider="openai" if "OPENAI_API_KEY" in os.environ else "hf"
-        # )
-
-        # Define tools for the agent
-        tools = [web_search, calculate]   # ← Add your custom tools here
-
         self.agent = CodeAgent(
             model=self.model,
-            tools=tools,
-            add_base_tools=True,          # includes Python interpreter, final_answer, etc.
+            tools=[web_search, calculate],
+            add_base_tools=True,      # includes final_answer tool etc.
             verbosity_level=1,
-            max_steps=12,                 # GAIA questions can need several steps
-            planning_interval=4           # helps with complex multi-step reasoning
+            max_steps=12,
+            planning_interval=4
         )
-        
         print("✅ Agent initialized successfully.")
 
     def __call__(self, question: str) -> str:
-        print(f"🤖 Agent processing question (first 80 chars): {question[:80]}...")
-        
+        print(f"🤖 Processing question: {question[:120]}...")
         try:
-            # Run the agent
             result = self.agent.run(question)
-            
-            # Smolagents usually returns the final answer nicely
             final_answer = str(result).strip()
-            
-            print(f"✅ Agent returned: {final_answer[:200]}{'...' if len(final_answer) > 200 else ''}")
+            print(f"✅ Answer generated ({len(final_answer)} chars)")
             return final_answer
-            
         except Exception as e:
             error_msg = f"Agent error: {str(e)}"
             print(f"❌ {error_msg}")
             return error_msg
 
 
-# ====================== THE REST OF THE CODE (unchanged) ======================
-# Keep everything from run_and_submit_all() onward exactly as you had it
-
+# ====================== RUN & SUBMIT ======================
 def run_and_submit_all(profile: gr.OAuthProfile | None):
-    """
-    Fetches all questions, runs the agent on them, submits all answers,
-    and displays the results.
-    """
     if not profile:
-        return "❌ Please Login to Hugging Face with the button.", None
+        return "❌ Please login with Hugging Face first.", None
 
     username = profile.username.strip()
-    print(f"👤 User logged in: {username}")
+    print(f"👤 Logged in as: {username}")
 
     api_url = DEFAULT_API_URL
     questions_url = f"{api_url}/questions"
     submit_url = f"{api_url}/submit"
 
-    # 1. Instantiate Agent
+    # Instantiate agent
     try:
         agent = BasicAgent()
     except Exception as e:
-        print(f"Error instantiating agent: {e}")
-        return f"Error initializing agent: {e}", None
+        return f"Error creating agent: {e}", None
 
     space_id = os.getenv("SPACE_ID")
     agent_code = f"https://huggingface.co/spaces/{space_id}/tree/main" if space_id else "https://huggingface.co/spaces/YOUR_USERNAME/YOUR_SPACE/tree/main"
 
-    # 2. Fetch Questions
+    # Fetch questions
     try:
         response = requests.get(questions_url, timeout=20)
         response.raise_for_status()
         questions_data = response.json()
-        if not questions_data:
-            return "Fetched questions list is empty.", None
         print(f"📥 Fetched {len(questions_data)} questions.")
     except Exception as e:
         return f"Error fetching questions: {e}", None
 
-    # 3. Run Agent on all questions
+    # Run agent on each question
     results_log = []
     answers_payload = []
 
-    print(f"🚀 Running agent on {len(questions_data)} questions... (this may take 5-15 minutes)")
-
     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:
+        question = item.get("question")
+        if not task_id or question is None:
             continue
 
         try:
-            submitted_answer = agent(question_text)
-            answers_payload.append({"task_id": task_id, "submitted_answer": submitted_answer})
+            answer = agent(question)
+            answers_payload.append({"task_id": task_id, "submitted_answer": answer})
             results_log.append({
                 "Task ID": task_id,
-                "Question": question_text[:150] + ("..." if len(question_text) > 150 else ""),
-                "Submitted Answer": str(submitted_answer)[:300] + ("..." if len(str(submitted_answer)) > 300 else "")
+                "Question": question[:150] + "..." if len(question) > 150 else question,
+                "Submitted Answer": str(answer)[:250] + "..." if len(str(answer)) > 250 else str(answer)
             })
         except Exception as e:
-            error_ans = f"AGENT ERROR: {str(e)}"
-            answers_payload.append({"task_id": task_id, "submitted_answer": error_ans})
-            results_log.append({"Task ID": task_id, "Question": question_text[:150]+"...", "Submitted Answer": error_ans})
+            err = f"ERROR: {str(e)}"
+            answers_payload.append({"task_id": task_id, "submitted_answer": err})
+            results_log.append({"Task ID": task_id, "Question": question[:150]+"...", "Submitted Answer": err})
 
     if not answers_payload:
-        return "No answers were generated.", pd.DataFrame(results_log)
+        return "No answers generated.", pd.DataFrame(results_log)
 
-    # 4. Submit
+    # Submit
     submission_data = {
         "username": username,
         "agent_code": agent_code,
@@ -183,17 +136,34 @@ def run_and_submit_all(profile: gr.OAuthProfile | None):
     try:
         response = requests.post(submit_url, json=submission_data, timeout=90)
         response.raise_for_status()
-        result_data = response.json()
-
-        final_status = (
-            f"✅ Submission Successful!\n\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\n"
-            f"Message: {result_data.get('message', 'No message')}"
-        )
+        result = response.json()
+
+        status = f"✅ Submission Successful!\n\n" \
+                 f"Score: {result.get('score', 'N/A')}% " \
+                 f"({result.get('correct_count', '?')}/{result.get('total_attempted', '?')})\n\n" \
+                 f"Message: {result.get('message', 'No message')}"
 
-        return final_status, pd.DataFrame(results_log)
+        return status, pd.DataFrame(results_log)
 
     except Exception as e:
-        return f"❌ Submission Failed: {str(e)}", pd.DataFrame(results_log)
+        return f"❌ Submission failed: {str(e)}", pd.DataFrame(results_log)
+
+
+# ====================== GRADIO UI ======================
+with gr.Blocks() as demo:
+    gr.Markdown("# 🤖 Unit 4 - GAIA Agent Evaluation")
+    gr.Markdown("Login → Click the button below. It may take 8–15 minutes to run all questions.")
+
+    gr.LoginButton()
+    run_button = gr.Button("🚀 Run Evaluation & Submit All Answers", variant="primary", size="large")
+
+    status_output = gr.Textbox(label="Status / Score", lines=10, interactive=False)
+    results_table = gr.DataFrame(label="Results", wrap=True)
+
+    run_button.click(
+        fn=run_and_submit_all,
+        outputs=[status_output, results_table]
+    )
+
+if __name__ == "__main__":
+    demo.launch(debug=True)