Update app.py

app.py

@@ -4,24 +4,29 @@ import requests
 import pandas as pd
 
 from smolagents import CodeAgent, DuckDuckGoSearchTool
-from smolagents.models import ApiModel
+from smolagents.models import InferenceClientModel
 
-# -----------------------------
+# ---------------------------------------------------
 # Constants
-# -----------------------------
+# ---------------------------------------------------
 DEFAULT_API_URL = "https://agents-course-unit4-scoring.hf.space"
 
-# -----------------------------
-# Build Agent
-# -----------------------------
+# ---------------------------------------------------
+# Agent Definition
+# ---------------------------------------------------
 class BasicAgent:
+
     def __init__(self):
 
-        # Uses your HF token automatically from Space secrets
-        model = ApiModel(
-            model_id="Qwen/Qwen2.5-72B-Instruct"
+        print("Initializing Hugging Face Agent...")
+
+        # Hugging Face Inference API model
+        model = InferenceClientModel(
+            model_id="Qwen/Qwen2.5-72B-Instruct",
+            token=os.getenv("HF_TOKEN")
         )
 
+        # Build agent
         self.agent = CodeAgent(
             tools=[
                 DuckDuckGoSearchTool()
@@ -31,79 +36,122 @@ class BasicAgent:
             verbosity_level=1
         )
 
-        print("HF Agent initialized.")
+        print("Agent initialized successfully.")
 
     def __call__(self, question: str) -> str:
 
+        print(f"\nQuestion: {question[:100]}")
+
         prompt = f"""
-You are a highly accurate GAIA benchmark solving agent.
+You are an expert GAIA benchmark solving agent.
 
-Rules:
-- Think carefully step by step.
-- Use web search when needed.
+Your job:
+- Think step-by-step.
+- Use web search if needed.
+- Solve the task accurately.
 - Return ONLY the final answer.
-- Keep the answer concise.
-- Do not explain reasoning unless asked.
+- Keep answers concise.
+- No explanations unless necessary.
 
 Question:
 {question}
 """
 
         try:
-            answer = self.agent.run(prompt)
 
-            if answer is None:
-                return "I could not determine the answer."
+            result = self.agent.run(prompt)
+
+            if result is None:
+                return "Could not determine the answer."
+
+            final_answer = str(result).strip()
 
-            return str(answer).strip()
+            print(f"Answer: {final_answer}")
+
+            return final_answer
 
         except Exception as e:
+
             print(f"Agent error: {e}")
-            return f"Error: {e}"
 
-# -----------------------------
-# Main Evaluation Function
-# -----------------------------
+            return f"Error: {str(e)}"
+
+
+# ---------------------------------------------------
+# Evaluation + Submission
+# ---------------------------------------------------
 def run_and_submit_all(profile: gr.OAuthProfile | None):
 
+    # Get Space ID
     space_id = os.getenv("SPACE_ID")
 
+    # Check login
     if profile:
-        username = f"{profile.username}"
-        print(f"User logged in: {username}")
+        username = profile.username
+        print(f"Logged in as: {username}")
     else:
-        return "Please login to Hugging Face.", None
+        return "Please login with Hugging Face first.", None
 
+    # API URLs
     api_url = DEFAULT_API_URL
     questions_url = f"{api_url}/questions"
     submit_url = f"{api_url}/submit"
 
+    # ---------------------------------------------------
     # Initialize Agent
+    # ---------------------------------------------------
     try:
         agent = BasicAgent()
+
     except Exception as e:
+
+        print(f"Error initializing agent: {e}")
+
         return f"Error initializing agent: {e}", None
 
+    # Link to your Space code
     agent_code = f"https://huggingface.co/spaces/{space_id}/tree/main"
 
+    # ---------------------------------------------------
     # Fetch Questions
+    # ---------------------------------------------------
     try:
-        response = requests.get(questions_url, timeout=30)
+
+        print("Fetching questions...")
+
+        response = requests.get(
+            questions_url,
+            timeout=30
+        )
+
         response.raise_for_status()
+
         questions_data = response.json()
 
+        print(f"Fetched {len(questions_data)} questions.")
+
     except Exception as e:
+
+        print(f"Error fetching questions: {e}")
+
         return f"Error fetching questions: {e}", None
 
+    # ---------------------------------------------------
     # Run Agent
-    results_log = []
+    # ---------------------------------------------------
     answers_payload = []
+    results_log = []
 
     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:
+            continue
+
+        print(f"\nRunning task: {task_id}")
+
         try:
 
             submitted_answer = agent(question_text)
@@ -121,19 +169,25 @@ def run_and_submit_all(profile: gr.OAuthProfile | None):
 
         except Exception as e:
 
+            print(f"Task error: {e}")
+
             results_log.append({
                 "Task ID": task_id,
                 "Question": question_text,
                 "Submitted Answer": f"ERROR: {e}"
             })
 
-    # Submit
+    # ---------------------------------------------------
+    # Submit Answers
+    # ---------------------------------------------------
     submission_data = {
         "username": username.strip(),
         "agent_code": agent_code,
         "answers": answers_payload
     }
 
+    print("Submitting answers...")
+
     try:
 
         response = requests.post(
@@ -147,45 +201,59 @@ def run_and_submit_all(profile: gr.OAuthProfile | None):
         result_data = response.json()
 
         final_status = (
-            f"Submission Successful!\n"
+            f"Submission Successful!\n\n"
             f"User: {result_data.get('username')}\n"
-            f"Overall Score: {result_data.get('score', 'N/A')}%\n"
-            f"Correct: {result_data.get('correct_count', '?')}/"
-            f"{result_data.get('total_attempted', '?')}\n"
+            f"Score: {result_data.get('score', 'N/A')}%\n"
+            f"Correct: "
+            f"{result_data.get('correct_count', '?')}/"
+            f"{result_data.get('total_attempted', '?')}\n\n"
             f"Message: {result_data.get('message', '')}"
         )
 
+        print(final_status)
+
         return final_status, pd.DataFrame(results_log)
 
     except Exception as e:
 
-        return (
-            f"Submission failed: {e}",
-            pd.DataFrame(results_log)
-        )
+        error_msg = f"Submission failed: {e}"
 
-# -----------------------------
-# Gradio UI
-# -----------------------------
+        print(error_msg)
+
+        return error_msg, pd.DataFrame(results_log)
+
+
+# ---------------------------------------------------
+# Gradio Interface
+# ---------------------------------------------------
 with gr.Blocks() as demo:
 
     gr.Markdown("# Hugging Face Agents Course - Final Assignment")
 
-    gr.Markdown("""
+    gr.Markdown(
+        """
 This agent uses:
+
 - Hugging Face Inference API
 - smolagents
 - DuckDuckGo web search
 - Qwen2.5-72B-Instruct
-""")
+"""
+    )
 
+    # HF Login
     gr.LoginButton()
 
-    run_button = gr.Button("Run Evaluation & Submit")
+    # Run Button
+    run_button = gr.Button(
+        "Run Evaluation & Submit"
+    )
 
+    # Outputs
     status_output = gr.Textbox(
         label="Submission Result",
-        lines=6
+        lines=8,
+        interactive=False
     )
 
     results_table = gr.DataFrame(
@@ -193,10 +261,25 @@ This agent uses:
         wrap=True
     )
 
+    # Button Action
     run_button.click(
         fn=run_and_submit_all,
-        outputs=[status_output, results_table]
+        outputs=[
+            status_output,
+            results_table
+        ]
     )
 
+# ---------------------------------------------------
+# Launch App
+# ---------------------------------------------------
 if __name__ == "__main__":
-    demo.launch(debug=True)
+
+    print("\n==============================")
+    print("Starting Hugging Face Agent...")
+    print("==============================\n")
+
+    demo.launch(
+        debug=True,
+        share=False
+    )