Update app.py

app.py

@@ -15,7 +15,6 @@ AGENT_LOAD_ERROR = ""
 AGENT_FUNCTION_NAME = "run_gaia_task" # Define the target function name
 
 try:
-    # --- MODIFIED: Import the correct function ---
     from final_agent import run_gaia_task
     print(f"Successfully imported {AGENT_FUNCTION_NAME} from final_agent.py")
     AGENT_AVAILABLE = True
@@ -24,15 +23,12 @@ except ImportError as e:
     print(error_msg)
     AGENT_LOAD_ERROR = error_msg
 except Exception as e:
-    # Catch errors during the global setup within final_agent.py
     error_msg = f"ERROR during import or initial setup in final_agent.py: {e}"
     print(error_msg)
     traceback.print_exc()
     AGENT_LOAD_ERROR = error_msg
 
-# --- MODIFIED: Define a dummy function matching the new signature ---
 if not AGENT_AVAILABLE:
-    # This dummy function will be used if the import fails
     def run_gaia_task(task_description: str) -> str:
         """Dummy function used when the real agent fails to load."""
         print(f"Executing dummy {AGENT_FUNCTION_NAME} because agent failed to load.")
@@ -44,32 +40,21 @@ class AgentRunner:
         print("AgentRunner initialized.")
         if not AGENT_AVAILABLE:
             print(f"WARNING: Agent function failed to load during startup. Error: {AGENT_LOAD_ERROR}")
-        # Optional: Add environment variable checks if needed
-        # if not os.getenv("GROQ_API_KEY") or not os.getenv("TAVILY_API_KEY"):
-        #     print("WARNING: Required API keys might not be set in Space secrets.")
 
     def __call__(self, question: str) -> str:
         """Runs the imported agent function on a single question."""
         print(f"\n--- AgentRunner received question: {question[:100]}... ---")
-        # Always call the potentially dummy function; it returns error if needed
         try:
-            # --- MODIFIED: Call the new agent function ---
-            # The 'question' variable holds the task description.
-            # The new agent handles file paths internally based on the description.
             final_answer = run_gaia_task(task_description=question)
-
-            # Ensure result is always a string for submission
             final_answer_str = str(final_answer)
             print(f"--- AgentRunner returning answer: {final_answer_str} ---")
             return final_answer_str
         except Exception as e:
-             # Catch unexpected errors during the function call itself
              print(f"!!! ERROR calling {AGENT_FUNCTION_NAME} function: {e} !!!")
-             traceback.print_exc() # Log the full error to Space logs
-             # --- MODIFIED: Update error message ---
+             traceback.print_exc()
              return f"ERROR: Agent function '{AGENT_FUNCTION_NAME}' failed during execution - {e}"
 
-# --- Submission Logic (Largely Unchanged) ---
+# --- Submission Logic ---
 def run_and_submit_all( profile: gr.OAuthProfile | None):
     """Fetches questions, runs agent, submits answers."""
     space_id = os.getenv("SPACE_ID")
@@ -82,9 +67,7 @@ def run_and_submit_all( profile: gr.OAuthProfile | None):
     # 1. Instantiate Agent Runner
     try:
         agent = AgentRunner()
-        # Check if agent loaded correctly before proceeding
         if not AGENT_AVAILABLE:
-             # --- MODIFIED: Update error message ---
              return f"Agent function '{AGENT_FUNCTION_NAME}' failed to load. Check logs. Error: {AGENT_LOAD_ERROR}", None
     except Exception as e: print(f"Error instantiating AgentRunner: {e}"); return f"Init error: {e}", None
 
@@ -106,18 +89,16 @@ def run_and_submit_all( profile: gr.OAuthProfile | None):
     question_count = len(questions_data)
     for i, item in enumerate(questions_data):
         task_id = item.get("task_id"); question_text = item.get("question")
-        print(f"\n--- Processing Question {i+1}/{question_count} (ID: {task_id}) ---") # Add progress logging
+        print(f"\n--- Processing Question {i+1}/{question_count} (ID: {task_id}) ---")
         if not task_id or question_text is None: print(f"Skipping item: {item}"); continue
         try:
-            # Calls AgentRunner.__call__, which now calls run_gaia_task
             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, "Submitted Answer": submitted_answer})
         except Exception as e:
-             # Catch errors during the agent's execution on a specific task
              print(f"!! Error running agent on task {task_id}: {e} !!"); traceback.print_exc()
              results_log.append({"Task ID": task_id, "Question": question_text, "Submitted Answer": f"AGENT RUN ERROR: {e}"})
-             answers_payload.append({"task_id": task_id, "submitted_answer": f"AGENT RUN ERROR: {e}"}) # Submit error
+             answers_payload.append({"task_id": task_id, "submitted_answer": f"AGENT RUN ERROR: {e}"})
 
     if not answers_payload: print("Agent produced no answers."); return "Agent produced no answers.", pd.DataFrame(results_log)
 
@@ -127,7 +108,7 @@ def run_and_submit_all( profile: gr.OAuthProfile | None):
 
     # 5. Submit
     try:
-        response = requests.post(submit_url, json=submission_data, timeout=120); response.raise_for_status() # Increased timeout
+        response = requests.post(submit_url, json=submission_data, timeout=120); response.raise_for_status()
         result_data = response.json()
         final_status = (f"Submission Successful!\nUser: {result_data.get('username')}\n"
                         f"Overall Score: {result_data.get('score', 'N/A')}% "
@@ -145,7 +126,7 @@ def run_and_submit_all( profile: gr.OAuthProfile | None):
     print(status_message); results_df = pd.DataFrame(results_log); return status_message, results_df
 
 
-# --- Build Gradio Interface (Unchanged) ---
+# --- Build Gradio Interface ---
 with gr.Blocks() as demo:
     gr.Markdown("# GAIA Agent Evaluation Runner")
     gr.Markdown(
@@ -158,19 +139,20 @@ with gr.Blocks() as demo:
         **Disclaimers:** Execution can take significant time depending on the number of questions and agent complexity.
         """
     )
-    gr.LoginButton()
+    login_button = gr.LoginButton() # Assign to variable to access profile info implicitly
     run_button = gr.Button("Run Evaluation & Submit All Answers")
     status_output = gr.Textbox(label="Run Status / Submission Result", lines=5, interactive=False)
     results_table = gr.DataFrame(label="Questions and Agent Answers", wrap=True)
-    run_button.click(fn=run_and_submit_all, inputs=gr.State(None), outputs=[status_output, results_table]) # Pass None for profile initially
 
-# --- Main execution block (Unchanged) ---
+    # --- CORRECTED LINE ---
+    # Remove the 'inputs' argument. The profile is passed implicitly because of LoginButton.
+    run_button.click(fn=run_and_submit_all, outputs=[status_output, results_table])
+
+# --- Main execution block ---
 if __name__ == "__main__":
     print("\n" + "-"*30 + " App Starting " + "-"*30)
-    # Perform startup checks if needed (e.g., check essential env vars)
     if not AGENT_AVAILABLE:
         print(f"CRITICAL WARNING: Agent function '{AGENT_FUNCTION_NAME}' could not be loaded. The app will run but agent calls will fail.")
         print(f"Load Error Details: {AGENT_LOAD_ERROR}")
     print("Launching Gradio Interface...")
-    # Consider removing debug=True for "production" submission space
     demo.launch(debug=False, share=False)