Update app.py

app.py

@@ -4,20 +4,24 @@ import os
 import gradio as gr
 import requests
 import pandas as pd
+import re
 
 from smolagents import CodeAgent, DuckDuckGoSearchTool
 from smolagents.models import OpenAIServerModel
 
-# Define the system prompt
 SYSTEM_PROMPT = """You are a general AI assistant. I will ask you a question.
-Report your thoughts, and finish your answer with the following template:
-FINAL ANSWER: [YOUR FINAL ANSWER].
-YOUR FINAL ANSWER should be a number OR as few words as possible OR a comma separated list 
-of numbers and/or strings. If you are asked for a number, don't use comma to write your number neither use units such as $ or percent sign unless specified otherwise. If you are asked for a string, don't use articles, neither abbreviations (e.g. for cities), and write the digits in plain text unless specified otherwise. If you are asked for a comma separated list, apply the above rules depending of whether the element to be put in the list is a number or a string."""
+Think step-by-step and finish your answer using the template:
+FINAL ANSWER: [YOUR FINAL ANSWER]
+
+Rules for FINAL ANSWER:
+- A number: no commas, units, or extra words. Use plain digits only.
+- A string: no articles or abbreviations. Use lowercase.
+- A list: comma-separated values, formatted as above.
+
+Only output the FINAL ANSWER line at the end. Do not explain the answer or repeat the question."""
 
 DEFAULT_API_URL = "https://agents-course-unit4-scoring.hf.space"
 
-# Patched model to prepend system prompt correctly
 class PatchedOpenAIServerModel(OpenAIServerModel):
     def generate(self, messages, stop_sequences=None, **kwargs):
         if isinstance(messages, list):
@@ -36,11 +40,21 @@ class MyAgent:
     def __call__(self, question: str) -> str:
         return self.agent.run(question)
 
+def extract_final_answer(output: str) -> str:
+    if "FINAL ANSWER:" in output:
+        return output.split("FINAL ANSWER:")[-1].strip().rstrip('.')
+    return output.strip()
+
+def sanitize_answer(ans: str) -> str:
+    ans = re.sub(r'\$|%|,', '', ans)
+    ans = ans.strip().rstrip('.')
+    return ans
+
 def run_and_submit_all(profile: gr.OAuthProfile | None):
     space_id = os.getenv("SPACE_ID")
 
     if profile:
-        username = profile.username
+        username = profile.username.strip()
         print(f"User logged in: {username}")
     else:
         print("User not logged in.")
@@ -79,7 +93,9 @@ def run_and_submit_all(profile: gr.OAuthProfile | None):
         if not task_id or question_text is None:
             continue
         try:
-            submitted_answer = agent(question_text)
+            raw_output = agent(question_text)
+            extracted = extract_final_answer(raw_output)
+            submitted_answer = sanitize_answer(extracted)
             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:
@@ -89,7 +105,7 @@ def run_and_submit_all(profile: gr.OAuthProfile | None):
     if not answers_payload:
         return "Agent did not produce any answers to submit.", pd.DataFrame(results_log)
 
-    submission_data = {"username": username.strip(), "agent_code": agent_code, "answers": answers_payload}
+    submission_data = {"username": username, "agent_code": agent_code, "answers": answers_payload}
     print(f"Submitting {len(answers_payload)} answers to: {submit_url}")
     try:
         response = requests.post(submit_url, json=submission_data, timeout=60)
@@ -115,7 +131,6 @@ def run_and_submit_all(profile: gr.OAuthProfile | None):
     except Exception as e:
         return f"An unexpected error occurred during submission: {e}", pd.DataFrame(results_log)
 
-# Gradio UI setup
 with gr.Blocks() as demo:
     gr.Markdown("# Basic Agent Evaluation Runner")
     gr.Markdown("""