Update app.py

app.py

@@ -9,7 +9,7 @@ from smolagents.models import InferenceClientModel
 # --- Constants ---
 DEFAULT_API_URL = "https://agents-course-unit4-scoring.hf.space"
 
-# --- Custom Tools with Proper Input/Output Specifications ---
+# --- Custom Tools ---
 class CalculatorTool(Tool):
     name = "calculator"
     description = "Performs mathematical calculations"
@@ -35,21 +35,21 @@ class CalculatorTool(Tool):
 class TimeTool(Tool):
     name = "current_time"
     description = "Gets current UTC time"
-    input_schema = {}  # No inputs needed
+    input_schema = {}  # No input
     output_schema = {
         "time": {
             "type": "string",
-            "description": "Current time in UTC (YYYY-MM-DD HH:MM:SS)"
+            "description": "Current UTC time (YYYY-MM-DD HH:MM:SS)"
         }
     }
 
     def use(self) -> dict:
         return {"time": datetime.utcnow().strftime("%Y-%m-%d %H:%M:%S UTC")}
 
-# --- Enhanced Agent ---
+# --- Agent Setup ---
 class LocalAgent:
     def __init__(self):
-        print("Initializing agent with smolagents...")
+        print("Initializing agent...")
         self.tools = [CalculatorTool(), TimeTool()]
         self.agent = ToolCallingAgent(
             tools=self.tools,
@@ -60,42 +60,32 @@ class LocalAgent:
         )
 
     def __call__(self, question: str) -> str:
-        print(f"Processing: {question[:100]}...")
         question_lower = question.lower()
+        if any(op in question_lower for op in ["calculate", "what is", "+", "-", "*", "/"]):
+            return CalculatorTool().use(question.replace("?", ""))["result"]
+        if "time" in question_lower:
+            return TimeTool().use()["time"]
 
-        # Direct tool usage for simple queries
-        if any(word in question_lower for word in ["calculate", "what is", "how much is", "+", "-", "*", "/"]):
-            result = CalculatorTool().use(question.replace("?", ""))
-            return result["result"]
-
-        if any(word in question_lower for word in ["time", "current time"]):
-            result = TimeTool().use()
-            return result["time"]
-
-        # Use full agent for complex questions
         try:
-            response = self.agent.run(question)
-            return str(response)
+            return str(self.agent.run(question))
         except Exception as e:
-            print(f"Agent error: {e}")
-            return "I couldn't process this question."
+            return f"Error: {e}"
 
-# --- Evaluation Runner ---
+# --- Evaluation Logic ---
 def run_and_submit_all(profile: gr.OAuthProfile | None):
     if not profile:
-        return "Please login first", None
+        return "Please login first.", None
 
     space_id = os.getenv("SPACE_ID", "local-test")
     api_url = os.getenv("API_URL", DEFAULT_API_URL)
 
     try:
         agent = LocalAgent()
-        response = requests.get(f"{api_url}/questions", timeout=15)
-        response.raise_for_status()
-        questions = response.json()
+        questions = requests.get(f"{api_url}/questions", timeout=15).json()
 
-        results = []
         answers = []
+        results = []
+
         for q in questions:
             try:
                 answer = agent(q["question"])
@@ -121,36 +111,34 @@ def run_and_submit_all(profile: gr.OAuthProfile | None):
             "answers": answers
         }
 
-        response = requests.post(f"{api_url}/submit", json=submission, timeout=60)
-        response.raise_for_status()
-        result = response.json()
+        result = requests.post(f"{api_url}/submit", json=submission, timeout=60).json()
 
         return (
-            f"Success! Score: {result.get('score', 'N/A')}%\n"
+            f"✅ Score: {result.get('score', 'N/A')}%\n"
             f"Correct: {result.get('correct_count', '?')}/{result.get('total_attempted', '?')}",
             pd.DataFrame(results)
         )
 
     except Exception as e:
-        return f"Evaluation failed: {str(e)}", pd.DataFrame(results if 'results' in locals() else [])
+        return f"Evaluation failed: {str(e)}", pd.DataFrame([])
 
-# --- Gradio Interface ---
+# --- Gradio UI ---
 with gr.Blocks(title="Agent Evaluation Runner") as app:
-    gr.Markdown("""
-    ## Advanced Agent Evaluation  
-    Uses smolagents with proper tool schemas
-    """)
+    gr.Markdown("## 🤖 Agent Evaluation with smolagents")
+    gr.Markdown("Login, then run the evaluation to test your agent.")
+
+    gr.LoginButton()  # OAuth Login UI
 
-    profile = gr.OAuthProfile()
-    run_btn = gr.Button("Run Evaluation")
-    output = gr.Textbox(label="Results")
-    results_table = gr.DataFrame(label="Question Log")
+    run_btn = gr.Button("🚀 Run Evaluation")
+    output = gr.Textbox(label="Result Summary")
+    results_table = gr.DataFrame(label="Answers and Logs")
 
     run_btn.click(
         fn=run_and_submit_all,
-        inputs=[profile],  # <-- Corrected: OAuthProfile passed in
+        inputs=[],  # No need to pass profile manually
         outputs=[output, results_table]
     )
 
 if __name__ == "__main__":
     app.launch()
+