Update app.py

app.py

@@ -1,44 +1,28 @@
 """
 app.py
 
-This script provides the Gradio web interface to run the evaluation for the
-Hugging Face Agents course. It fetches questions from a server, runs the
-agent defined in agent.py on them, and submits the answers for scoring.
+This script provides the Gradio web interface to run the evaluation.
+This version has been corrected to be "file-aware" by checking for a 'file_url'
+in the task data and appending it to the agent's prompt.
 """
 
 import os
-import re # <-- 1. ADDED IMPORT for Regular Expressions
+import re
 import gradio as gr
 import requests
 import pandas as pd
 
-# --- Import your agent's factory function ---
-from agent import create_agent_executor # <-- 2. ADDED IMPORT for your agent
+from agent import create_agent_executor
 
 # --- Constants ---
 DEFAULT_API_URL = "https://agents-course-unit4-scoring.hf.space"
 
-# --- DELETED BasicAgent class as it's no longer needed ---
-
-# --- 3. ADDED HELPER FUNCTION TO PARSE THE AGENT'S OUTPUT ---
+# --- Helper function to parse the agent's output (This is correct) ---
 def parse_final_answer(agent_response: str) -> str:
-    """
-    Extracts the final answer from the agent's full response string.
-    The agent is prompted to return 'FINAL ANSWER: [answer]'.
-    This function isolates and returns '[answer]'.
-    """
-    # Use a regular expression to find the text after "FINAL ANSWER:"
     match = re.search(r"FINAL ANSWER:\s*(.*)", agent_response, re.IGNORECASE | re.DOTALL)
-    if match:
-        # If a match is found, return the captured group, stripped of whitespace
-        return match.group(1).strip()
-    
-    # As a fallback, if the specific format is not found, return the last non-empty line
+    if match: return match.group(1).strip()
     lines = [line for line in agent_response.split('\n') if line.strip()]
-    if lines:
-        return lines[-1].strip()
-        
-    # If all else fails, return a default message
+    if lines: return lines[-1].strip()
     return "Could not parse a final answer."
 
 
@@ -48,7 +32,6 @@ def run_and_submit_all(profile: gr.OAuthProfile | None):
     and displays the results.
     """
     if not profile:
-        print("User not logged in.")
         return "Please log in to Hugging Face with the button above to submit.", None
     
     username = profile.username
@@ -58,18 +41,15 @@ def run_and_submit_all(profile: gr.OAuthProfile | None):
     agent_code = f"https://huggingface.co/spaces/{space_id}/tree/main"
     questions_url = f"{DEFAULT_API_URL}/questions"
     submit_url = f"{DEFAULT_API_URL}/submit"
-
-    # --- 4. MODIFIED AGENT INSTANTIATION AND EXECUTION ---
     
-    # 1. Instantiate Agent (using your factory function from agent.py)
+    # 1. Instantiate Agent
     print("Initializing your custom agent...")
     try:
-        agent_executor = create_agent_executor(provider="groq") # or "groq"
+        agent_executor = create_agent_executor(provider="google") # Using Google for better tool use
     except Exception as e:
-        print(f"Error instantiating agent: {e}")
         return f"Fatal Error: Could not initialize agent. Check logs. Details: {e}", None
 
-    # 2. Fetch Questions (this part is correct)
+    # 2. Fetch Questions
     print(f"Fetching questions from: {questions_url}")
     try:
         response = requests.get(questions_url, timeout=20)
@@ -77,32 +57,37 @@ def run_and_submit_all(profile: gr.OAuthProfile | None):
         questions_data = response.json()
         print(f"Fetched {len(questions_data)} questions.")
     except Exception as e:
-        return f"Error fetching questions: {e}", None
+        return f"Error fetching questions: {e}", pd.DataFrame()
 
-    # 3. Run your Agent (THIS IS THE MOST IMPORTANTLY CORRECTED SECTION)
-    results_log = []
-    answers_payload = []
+    # 3. Run your Agent
+    results_log, answers_payload = [], []
     print(f"Running agent on {len(questions_data)} questions...")
     for i, item in enumerate(questions_data):
         task_id = item.get("task_id")
         question_text = item.get("question")
-        
-        if not task_id or question_text is None:
-            print(f"Skipping item with missing data: {item}")
-            continue
+        if not task_id or question_text is None: continue
         
         print(f"\n--- Running Task {i+1}/{len(questions_data)} (ID: {task_id}) ---")
-        print(f"Question: {question_text}")
         
+        # --- THIS IS THE CRITICAL FIX ---
+        # 1. Check if a 'file_url' key exists in the task data.
+        file_url = item.get("file_url")
+        full_question_text = question_text
+
+        # 2. If a URL exists, append it to the question text.
+        if file_url:
+            print(f"File found for this task: {file_url}")
+            # This gives the agent the context it needs to call the right tool.
+            full_question_text = f"{question_text}\n\n[Attachment URL: {file_url}]"
+        
+        print(f"Full Prompt for Agent:\n{full_question_text}")
+        # --- END CRITICAL FIX ---
+
         try:
-            # CORRECT INVOCATION: Use the agent_executor with the .invoke() method
-            # The input must be a dictionary with a "messages" key
-            result = agent_executor.invoke({"messages": [("user", question_text)]})
+            # 3. Pass the full, potentially enriched, question text to the agent.
+            result = agent_executor.invoke({"messages": [("user", full_question_text)]})
             
-            # The agent's final response is in the 'messages' list of the output
             raw_answer = result['messages'][-1].content
-            
-            # Use our helper function to extract the clean answer
             submitted_answer = parse_final_answer(raw_answer)
             
             print(f"Raw LLM Response: '{raw_answer}'")
@@ -112,37 +97,29 @@ def run_and_submit_all(profile: gr.OAuthProfile | None):
             results_log.append({"Task ID": task_id, "Question": question_text, "Submitted Answer": submitted_answer})
         except Exception as e:
              print(f"!! AGENT ERROR on task {task_id}: {e}")
-             # It's important to log errors so you can see them in the UI
              results_log.append({"Task ID": task_id, "Question": question_text, "Submitted Answer": f"AGENT RUNTIME ERROR: {e}"})
 
     if not answers_payload:
         return "Agent did not produce any answers to submit.", pd.DataFrame(results_log)
 
-    # 4. Prepare and 5. Submit (these parts are correct)
+    # 4. Prepare and 5. Submit
     submission_data = {"username": username, "agent_code": agent_code, "answers": answers_payload}
     print(f"\nSubmitting {len(answers_payload)} answers for user '{username}'...")
     try:
         response = requests.post(submit_url, json=submission_data, timeout=60)
         response.raise_for_status()
         result_data = response.json()
-        final_status = (
-            f"Submission Successful!\n"
-            f"User: {result_data.get('username')}\n"
-            f"Overall Score: {result_data.get('score', 'N/A')}% "
-            f"({result_data.get('correct_count', '?')}/{result_data.get('total_attempted', '?')} correct)"
-        )
-        print("Submission successful.")
+        final_status = (f"Submission Successful!\nUser: {result_data.get('username')}\nOverall Score: {result_data.get('score', 'N/A')}%")
         return final_status, pd.DataFrame(results_log)
     except Exception as e:
         status_message = f"Submission Failed: {e}"
         print(status_message)
         return status_message, pd.DataFrame(results_log)
 
-
-# --- Build Gradio Interface using Blocks (This part is correct) ---
+# --- Gradio UI (This part is correct) ---
 with gr.Blocks() as demo:
     gr.Markdown("# Agent Evaluation Runner")
-    gr.Markdown("Run your custom agent against the evaluation questions and submit for a score.")
+    # ... (rest of the Gradio code is fine)
     gr.LoginButton()
     run_button = gr.Button("Run Evaluation & Submit All Answers", variant="primary")
     status_output = gr.Textbox(label="Run Status / Submission Result", lines=4, interactive=False)
@@ -151,13 +128,4 @@ with gr.Blocks() as demo:
 
 if __name__ == "__main__":
     print("\n" + "-"*30 + " App Starting " + "-"*30)
-    # The startup info logs are helpful, so we keep them.
-    space_id_startup = os.getenv("SPACE_ID")
-    if space_id_startup:
-        print(f"✅ SPACE_ID found: {space_id_startup}")
-        print(f"   Repo URL: https://huggingface.co/spaces/{space_id_startup}")
-    else:
-        print("ℹ️  SPACE_ID environment variable not found (likely running locally).")
-    print("-"*(60 + len(" App Starting ")) + "\n")
-    print("Launching Gradio Interface...")
     demo.launch()