Update app.py

app.py

@@ -4,7 +4,6 @@ from openai import OpenAI
 import gradio as gr
 import requests
 import pandas as pd
-
 from smolagents import CodeAgent, DuckDuckGoSearchTool, tool
 
 # --- Constants ---
@@ -16,20 +15,27 @@ if not openai_api_key:
     raise RuntimeError("Please set OPENAI_API_KEY in your Space secrets or env!")
 openai.api_key = openai_api_key
 client = OpenAI()
-
 OPENAI_MODEL_ID = os.getenv("OPENAI_MODEL_ID", "gpt-4.1")
 
-# --- Tool Definitions ---
+# --- Model Wrapper for CodeAgent compatibility ---
+class OpenAIModelWrapper:
+    def __init__(self, model_id: str, client: OpenAI):
+        self.model_id = model_id
+        self.client = client
+
+    def __call__(self, prompt: str, **kwargs) -> str:
+        # ignore extra kwargs (stop_sequences, temperature, etc.)
+        resp = self.client.responses.create(
+            model=self.model_id,
+            input=prompt
+        )
+        return resp.output_text
 
+# --- Tool Definitions ---
 @tool
 def summarize_query(query: str) -> str:
     """
-    Reframes an unclear query into a better one.
-
-    Args:
-        query (str): The search query to refine.
-    Returns:
-        str: A concise, improved query.
+    Reframes an unclear query into a concise one.
     """
     return f"Summarize and reframe: {query}"
 
@@ -37,25 +43,19 @@ def summarize_query(query: str) -> str:
 def wikipedia_search(page: str) -> str:
     """
     Fetches the summary extract of an English Wikipedia page.
-
-    Args:
-        page (str): The page title (e.g. 'Mercedes_Sosa_discography').
-    Returns:
-        str: The extract section text.
     """
     url = f"https://en.wikipedia.org/api/rest_v1/page/summary/{page}"
-    resp = requests.get(url, timeout=10)
-    resp.raise_for_status()
-    return resp.json().get("extract", "")
+    r = requests.get(url, timeout=10)
+    r.raise_for_status()
+    return r.json().get("extract", "")
 
 search_tool    = DuckDuckGoSearchTool()
 wiki_tool      = wikipedia_search
 summarize_tool = summarize_query
 
 # --- ReACT + Scratchpad + Auto-Retry Instruction Prompt ---
-
 instruction_prompt = """
-You are a ReACT agent with three tools:
+You are a ReACT agent with three tools: 
  • DuckDuckGoSearchTool(query: str)
  • wikipedia_search(page: str)
  • summarize_query(query: str)
@@ -69,8 +69,8 @@ Internally, for each question:
    Record new Observation.
 5. Thought: integrate observations.
 
-Finally, output exactly one line:
-FINAL ANSWER: [your concise answer]
+Finally, output 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.
+
 
 Rules:
 - Numbers: digits only.
@@ -78,63 +78,43 @@ Rules:
 - Strings: no filler words.
 """
 
-# --- Model wrapper to satisfy CodeAgent's expected interface ---
-
-class OpenAIModelWrapper:
-    def __init__(self, model_id: str, client: OpenAI):
-        self.model_id = model_id
-        self.client = client
-
-    def __call__(self, prompt: str, **kwargs) -> str:
-        # kwargs catches stop_sequences, temperature, etc.
-        resp = self.client.responses.create(
-            model=self.model_id,
-            input_text=prompt       # <-- use input_text here
-        )
-        return resp.output_text
-
-llm_wrapper = OpenAIModelWrapper(model_id=OPENAI_MODEL_ID, client=client)
-
 # --- Build the CodeAgent ---
-
+llm_wrapper = OpenAIModelWrapper(model_id=OPENAI_MODEL_ID, client=client)
 smart_agent = CodeAgent(
     tools=[search_tool, wiki_tool, summarize_tool],
     model=llm_wrapper
 )
 
 # --- BasicAgent wrapper for Gradio ---
-
 class BasicAgent:
     def __init__(self):
         print("SmolAgent (GPT-4.1) with ReACT & tools initialized.")
 
     def __call__(self, question: str) -> str:
         prompt = instruction_prompt.strip() + "\n\nQUESTION: " + question.strip()
-        print(f"Agent prompt: {prompt[:120]}…")
         try:
             return smart_agent.run(prompt)
         except Exception as e:
             return f"AGENT ERROR: {e}"
 
 # --- Gradio / Submission Logic ---
-
 def run_and_submit_all(profile: gr.OAuthProfile | None):
     if not profile:
         return "Please log in to Hugging Face.", None
     username   = profile.username
     space_id   = os.getenv("SPACE_ID", "")
-    agent      = BasicAgent()
     agent_code = f"https://huggingface.co/spaces/{space_id}/tree/main"
+    agent      = BasicAgent()
 
-    # 1. Fetch questions
+    # Fetch questions
     try:
-        resp      = requests.get(f"{DEFAULT_API_URL}/questions", timeout=15)
+        resp = requests.get(f"{DEFAULT_API_URL}/questions", timeout=15)
         resp.raise_for_status()
         questions = resp.json() or []
     except Exception as e:
         return f"Error fetching questions: {e}", None
 
-    # 2. Run agent on each question
+    # Run agent
     logs, payload = [], []
     for item in questions:
         tid = item.get("task_id")
@@ -148,7 +128,7 @@ def run_and_submit_all(profile: gr.OAuthProfile | None):
     if not payload:
         return "Agent did not produce any answers.", pd.DataFrame(logs)
 
-    # 3. Submit answers
+    # Submit
     submission = {"username": username, "agent_code": agent_code, "answers": payload}
     try:
         post = requests.post(f"{DEFAULT_API_URL}/submit", json=submission, timeout=60)
@@ -157,8 +137,8 @@ def run_and_submit_all(profile: gr.OAuthProfile | None):
         status = (
             f"Submission Successful!\n"
             f"User: {res.get('username')}\n"
-            f"Overall Score: {res.get('score', 'N/A')}% "
-            f"({res.get('correct_count', '?')}/{res.get('total_attempted', '?')})\n"
+            f"Overall Score: {res.get('score','N/A')}% "
+            f"({res.get('correct_count','?')}/{res.get('total_attempted','?')})\n"
             f"Message: {res.get('message','')}"
         )
         return status, pd.DataFrame(logs)
@@ -166,7 +146,6 @@ def run_and_submit_all(profile: gr.OAuthProfile | None):
         return f"Submission Failed: {e}", pd.DataFrame(logs)
 
 # --- Gradio Interface ---
-
 with gr.Blocks() as demo:
     gr.Markdown("# SmolAgent GAIA Runner 🚀")
     gr.Markdown("""
@@ -180,7 +159,6 @@ with gr.Blocks() as demo:
     run_btn    = gr.Button("Run Evaluation & Submit All Answers")
     status_out = gr.Textbox(label="Status", lines=5, interactive=False)
     table_out  = gr.DataFrame(label="Questions & Answers", wrap=True)
-
     run_btn.click(fn=run_and_submit_all, outputs=[status_out, table_out])
 
 if __name__ == "__main__":