test4

app.py

@@ -3,11 +3,11 @@ import gradio as gr
 import requests
 import openai
 
-# Setup OpenAI API Key dari Environment
+# Setup OpenAI API Key
 openai.api_key = os.getenv("OPENAI_API_KEY")
 DEFAULT_API_URL = "https://agents-course-unit4-scoring.hf.space"
 
-# ----------- AGENT GPT -----------
+# Agent GPT
 class GPTAgent:
     def __init__(self, model="gpt-4"):
         self.model = model
@@ -19,7 +19,7 @@ class GPTAgent:
                 messages=[
                     {
                         "role": "system",
-                        "content": "You are a helpful assistant that answers only with one of the letters: A, B, C, or D.",
+                        "content": "You are a helpful assistant that only replies with A, B, C, or D.",
                     },
                     {"role": "user", "content": question},
                 ],
@@ -30,30 +30,30 @@ class GPTAgent:
         except Exception as e:
             return f"ERROR: {str(e)}"
 
-# ----------- Main Logic -----------
-def run_and_submit_all(profile):
-    username = getattr(profile, "username", "anonymous")
+# Evaluation Function
+def run_and_submit_all(user_profile):
+    username = getattr(user_profile, "username", "anonymous")
     space_id = os.getenv("SPACE_ID", "anonymous/Final_Assignment_Template")
 
     try:
         agent = GPTAgent()
     except Exception as e:
-        return f"❌ Agent init error: {e}", ""
+        return f"❌ Agent error: {e}", ""
 
     try:
         questions = requests.get(f"{DEFAULT_API_URL}/questions", timeout=10).json()
     except Exception as e:
-        return f"❌ Failed to fetch questions: {e}", ""
+        return f"❌ Could not fetch questions: {e}", ""
 
     answers = []
     logs = []
 
-    for i, q in enumerate(questions, 1):
+    for idx, q in enumerate(questions, 1):
         task_id = q.get("task_id")
         question = q.get("question")
         answer = agent(question)
         answers.append({"task_id": task_id, "answer": answer})
-        logs.append(f"### Q{i} - {task_id}\n{question}\n**Answer:** {answer}\n")
+        logs.append(f"### Q{idx} - {task_id}\n{question}\n**Answer:** {answer}\n")
 
     try:
         result = requests.post(
@@ -75,16 +75,16 @@ def run_and_submit_all(profile):
 
     return summary, "\n\n".join(logs)
 
-# ------------- UI ---------------
+# Gradio UI
 with gr.Blocks() as demo:
-    gr.Markdown("# 🤖 GAIA Evaluation Agent with GPT\nLogin with Hugging Face to evaluate your agent.")
+    gr.Markdown("# 🧠 GAIA Benchmark Agent\nLogin with your Hugging Face account to begin.")
 
-    profile = gr.OAuthProfile()
-    output_summary = gr.Textbox(label="📊 Result", lines=2)
-    output_log = gr.Textbox(label="📋 Answer Log", lines=20, show_copy_button=True)
+    login_btn = gr.LoginButton()
+    profile_output = gr.OAuthProfile()
+    run_button = gr.Button("▶ Run Evaluation")
+    summary = gr.Textbox(label="📊 Result", lines=2)
+    log_output = gr.Textbox(label="📋 Log", lines=20, show_copy_button=True)
 
-    run_button = gr.Button("▶ Run All and Submit")
+    run_button.click(fn=run_and_submit_all, inputs=[profile_output], outputs=[summary, log_output])
 
-    run_button.click(fn=run_and_submit_all, inputs=[profile], outputs=[output_summary, output_log])
-
-demo.launch()
+demo.launch()