remove final answer

agent.py

@@ -19,7 +19,6 @@ from supabase.client import Client, create_client
 load_dotenv()
 
 # === Tools ===
-
 @tool
 def multiply(a: int, b: int) -> int:
     """Multiply two integers."""
@@ -72,7 +71,7 @@ sys_msg = SystemMessage(content=system_prompt)
 
 # === Embeddings and Vector Store ===
 embeddings = HuggingFaceEmbeddings(model_name="sentence-transformers/all-mpnet-base-v2")
-supabase: Client = create_client(os.environ.get("SUPABASE_URL"), os.environ.get("SUPABASE_SERVICE_KEY"))
+supabase = create_client(os.getenv("SUPABASE_URL"), os.getenv("SUPABASE_SERVICE_KEY"))
 vector_store = SupabaseVectorStore(
     client=supabase,
     embedding=embeddings,
@@ -83,7 +82,7 @@ vector_store = SupabaseVectorStore(
 # === Tools ===
 tools = [multiply, add, subtract, divide, modulus, wiki_search, web_search, arvix_search]
 
-# === Build Graph ===
+# === Graph Builder ===
 def build_graph(provider: str = "groq"):
     if provider == "google":
         llm = ChatGoogleGenerativeAI(model="gemini-2.0-flash", temperature=0)
@@ -104,6 +103,7 @@ def build_graph(provider: str = "groq"):
     def assistant(state: MessagesState):
         response = llm_with_tools.invoke(state["messages"])
         content = response.content.strip()
+        # Extract exact match content, remove FINAL ANSWER: if present
         if "FINAL ANSWER:" in content:
             content = content.split("FINAL ANSWER:")[-1].strip()
         return {"messages": [AIMessage(content=content)]}

app.py

@@ -20,25 +20,20 @@ DEFAULT_API_URL = "https://agents-course-unit4-scoring.hf.space"
 
 cached_answers = []
 
+def parse_final_answer(raw: str) -> str:
+    raw = raw.strip()
+    if "FINAL ANSWER:" in raw:
+        return raw.split("FINAL ANSWER:")[-1].strip()
+    return raw.split("Final Answer:")[-1].strip() if "Final Answer:" in raw else raw
+
 class BasicAgent:
-    """A langgraph agent."""
     def __init__(self):
-        print("BasicAgent initialized.")
         self.graph = build_graph()
 
     def __call__(self, question: str) -> str:
-        print(f"Agent received question (first 50 chars): {question[:50]}...")
         messages = [HumanMessage(content=question)]
-        result = self.graph.invoke({"messages": messages})
-        full_output = result["messages"][-1].content.strip()
-
-        # Enforce strict FINAL ANSWER format parsing
-        match = re.search(r"FINAL ANSWER:\s*(.+)", full_output, re.IGNORECASE)
-        if match:
-            return match.group(0).strip()  # Returns the entire "FINAL ANSWER: xxx"
-        else:
-            print(" FINAL ANSWER not found in output, returning fallback.")
-            return "FINAL ANSWER: unknown"
+        output = self.graph.invoke({"messages": messages})
+        return parse_final_answer(output['messages'][-1].content)
 
 def run_agent_only(profile: gr.OAuthProfile | None):
     global cached_answers
@@ -53,86 +48,66 @@ def run_agent_only(profile: gr.OAuthProfile | None):
     except Exception as e:
         return f"Agent Init Error: {e}", None
 
-    questions_url = f"{DEFAULT_API_URL}/questions"
-
     try:
-        response = requests.get(questions_url, timeout=15)
-        questions_data = response.json()
+        questions = requests.get("https://agents-course-unit4-scoring.hf.space/questions", timeout=15).json()
     except Exception as e:
         return f"Error fetching questions: {e}", None
 
-    with open("system_prompt.txt", "r", encoding="utf-8") as f:
+    with open("system_prompt.txt", "r") as f:
         system_prompt = f.read().strip()
 
-    for item in questions_data:
+    for item in questions:
         task_id = item.get("task_id")
         question = item.get("question")
         file_name = item.get("file_name")
 
-        if not task_id or question is None:
+        if not task_id or not question:
             continue
 
         try:
-            user_message = question
-            if file_name:
-                user_message += f"\n\nFile to use: {file_name}"
-
-            full_input = system_prompt + "\n\n" + user_message
-            answer = agent(full_input)
+            user_message = question + (f"\n\nFile to use: {file_name}" if file_name else "")
+            answer = agent(system_prompt + "\n\n" + user_message)
             cached_answers.append({"task_id": task_id, "submitted_answer": answer})
             results_log.append({"Task ID": task_id, "Question": question, "Submitted Answer": answer})
         except Exception as e:
             results_log.append({"Task ID": task_id, "Question": question, "Submitted Answer": f"AGENT ERROR: {e}"})
 
-    return "Agent finished. Now click 'Submit Cached Answers'", pd.DataFrame(results_log)
+    return "Agent finished. Click 'Submit Cached Answers' next.", pd.DataFrame(results_log)
 
 def submit_cached_answers(profile: gr.OAuthProfile | None):
-    global cached_answers
     if not profile or not cached_answers:
-        return "No cached answers to submit. Run the agent first.", None
-
-    space_id = os.getenv("SPACE_ID")
-    username = profile.username
-    agent_code = f"https://huggingface.co/spaces/{space_id}/tree/main"
+        return "Nothing to submit. Run the agent first.", None
 
     payload = {
-        "username": username,
-        "agent_code": agent_code,
+        "username": profile.username,
+        "agent_code": f"https://huggingface.co/spaces/{os.getenv('SPACE_ID')}/tree/main",
         "answers": cached_answers
     }
 
-    submit_url = f"{DEFAULT_API_URL}/submit"
-
     try:
-        response = requests.post(submit_url, json=payload, timeout=60)
+        response = requests.post("https://agents-course-unit4-scoring.hf.space/submit", json=payload, timeout=60)
         result = response.json()
-        final_status = (
-            f"Submission Successful!\nUser: {result.get('username')}\n"
-            f"Score: {result.get('score', 'N/A')}% ({result.get('correct_count', '?')}/{result.get('total_attempted', '?')})"
-        )
-        return final_status, None
+        score = result.get("score", "?")
+        correct = result.get("correct_count", "?")
+        total = result.get("total_attempted", "?")
+        return f"Submission complete. Score: {score}% ({correct}/{total})", None
     except Exception as e:
         return f"Submission failed: {e}", None
 
-# --- Gradio UI ---
 with gr.Blocks() as demo:
-    gr.Markdown("# Basic Agent Evaluation Runner")
-    gr.Markdown("""
-    **Instructions:**
-    1. Run the Agent to generate answers to all questions.
-    2. Then click 'Submit Cached Answers' to submit them for scoring.
-    """)
+    gr.Markdown("""# Agent Evaluator
+1. Login with Hugging Face
+2. Run agent only
+3. Submit answers""")
 
     gr.LoginButton()
+    run_button = gr.Button("Run Agent")
+    submit_button = gr.Button("Submit Cached Answers")
+    status_box = gr.Textbox(label="Status", lines=4)
+    table = gr.DataFrame(label="Answers Log")
 
-    run_button = gr.Button("🧠 Run Agent Only")
-    submit_button = gr.Button("📤 Submit Cached 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_agent_only, outputs=[status_output, results_table])
-    submit_button.click(fn=submit_cached_answers, outputs=[status_output, results_table])
+    run_button.click(fn=run_agent_only, outputs=[status_box, table])
+    submit_button.click(fn=submit_cached_answers, outputs=[status_box, table]))
 
 if __name__ == "__main__":
     print("\n" + "-"*30 + " App Starting " + "-"*30)