test4

app.py

@@ -3,73 +3,66 @@ import gradio as gr
 import requests
 import openai
 
-# === Konfigurasi API ===
-DEFAULT_API_URL = "https://agents-course-unit4-scoring.hf.space"
-openai.api_key = os.getenv("OPENAI_API_KEY")
-
-# === Agent GPT-4o ===
+# Gunakan OpenAI GPT untuk menjawab pertanyaan
 class GPTAgent:
     def __init__(self):
-        print("✅ GPTAgent initialized")
+        self.api_key = os.getenv("OPENAI_API_KEY")
+        openai.api_key = self.api_key
+        self.model = "gpt-4o"  # atau "gpt-3.5-turbo" jika tidak punya akses
 
     def __call__(self, question: str) -> str:
         try:
             response = openai.chat.completions.create(
-                model="gpt-4o",
+                model=self.model,
                 messages=[
-                    {"role": "system", "content": "You are an expert AI assistant."},
-                    {"role": "user", "content": question}
-                ],
-                temperature=0.2,
-                max_tokens=500,
+                    {"role": "system", "content": "You are a helpful assistant."},
+                    {"role": "user", "content": question},
+                ]
             )
             return response.choices[0].message.content.strip()
         except Exception as e:
             return f"ERROR: {e}"
 
-# === Fungsi Evaluasi & Submit ===
+# Fungsi utama untuk evaluasi & submit
 def run_and_submit_all(user_profile):
-    username = user_profile.get("username", "anonymous")
-    questions_url = f"{DEFAULT_API_URL}/questions"
-    submit_url = f"{DEFAULT_API_URL}/submit"
-
+    username = user_profile if user_profile else "anonymous"
     agent = GPTAgent()
 
+    questions_url = "https://agents-course-unit4-scoring.hf.space/questions"
+    submit_url = "https://agents-course-unit4-scoring.hf.space/submit"
+
     try:
-        response = requests.get(questions_url, timeout=10)
-        response.raise_for_status()
-        questions = response.json()
+        questions = requests.get(questions_url, timeout=10).json()
     except Exception as e:
-        return f"Error fetching questions: {e}", []
+        return f"❌ Error fetching questions: {e}", []
 
+    answers = []
     results = []
-    for item in questions:
-        task_id = item.get("task_id", "")
-        question = item.get("question", "")
-        answer = agent(question)
-        results.append([task_id, question, answer])
 
-    # Format payload sesuai API GAIA
+    for q in questions:
+        task_id = q["task_id"]
+        question_text = q["question"]
+        answer = agent(question_text)
+        answers.append({"task_id": task_id, "answer": answer})
+        results.append([task_id, question_text, answer])
+
     payload = {
         "user": username,
-        "answers": [{"task_id": r[0], "answer": r[2]} for r in results],
+        "answers": answers
     }
 
     try:
-        submit_resp = requests.post(submit_url, json=payload, timeout=10)
-        submit_resp.raise_for_status()
-        score_data = submit_resp.json()
-        status = (
-            f"User: {username}\n"
-            f"Score: {score_data.get('score', 0.0) * 100:.1f}%\n"
-            f"Correct: {score_data.get('correct', 0)}/20"
-        )
+        res = requests.post(submit_url, json=payload, timeout=10)
+        res.raise_for_status()
+        score = res.json().get("score", 0.0)
+        correct = res.json().get("correct", 0)
+        total = res.json().get("total", len(questions))
+        status = f"✅ User: {username}\nScore: {score:.1f}%\nCorrect: {correct}/{total}"
+        return status, results
     except Exception as e:
-        return f"Error submitting answers: {e}\nResponse: {submit_resp.text}", results
-
-    return status, results
+        return f"❌ Error submitting answers: {e}", results
 
-# === Tampilan antarmuka Gradio ===
+# UI Gradio
 with gr.Blocks() as demo:
     gr.Markdown("# GAIA Benchmark Agent with GPT-4o")
     gr.Markdown(
@@ -80,12 +73,12 @@ with gr.Blocks() as demo:
         """
     )
 
-    login_btn = gr.LoginButton()
+    login = gr.LoginButton()
     run_btn = gr.Button("Run Evaluation & Submit")
     status = gr.Textbox(label="Status", interactive=False, lines=4)
-    results = gr.DataFrame(headers=["Task ID", "Question", "Answer"], interactive=False)
+    results = gr.DataFrame(headers=["Task ID", "Question", "Answer"])
 
-    run_btn.click(fn=run_and_submit_all, inputs=[login_btn], outputs=[status, results])
+    run_btn.click(fn=run_and_submit_all, inputs=[login], outputs=[status, results])
 
 if __name__ == "__main__":
     demo.launch()