Update app.py

app.py

@@ -3,8 +3,9 @@ import gradio as gr
 import requests
 import inspect
 import pandas as pd
-from smolagents import CodeAgent, DuckDuckGoSearchTool, InferenceClientModel
+from smolagents import CodeAgent, DuckDuckGoSearchTool, HfApiModel
 
+# (Keep Constants and BasicAgent class as is)
 # --- Constants ---
 DEFAULT_API_URL = "https://agents-course-unit4-scoring.hf.space"
 
@@ -12,52 +13,23 @@ DEFAULT_API_URL = "https://agents-course-unit4-scoring.hf.space"
 class BasicAgent:
     def __init__(self):
         print("BasicAgent initialized.")
-
-        # --- CHOOSE THE BEST MODEL HERE ---
-        # Using Meta-Llama-3-8B-Instruct for a good balance of performance and cost.
-        # Ensure you have accepted its terms on Hugging Face and have sufficient
-        # Inference Provider credits or a PRO subscription with pay-as-you-go enabled.
-        self.agent = CodeAgent(
-            tools=[DuckDuckGoSearchTool()],
-            model=InferenceClientModel(model_id="meta-llama/Meta-Llama-3-8B-Instruct"),
-            max_steps=10, # Increase max_steps for more complex reasoning chains if needed
-            # use_structured_outputs_internally=True # Optional: Can sometimes improve performance for capable models, but might also cause issues if the model doesn't strictly adhere. Test this if needed.
-        )
-
-        # --- REVISED SYSTEM PROMPT FOR EXACT MATCH ---
-        # This prompt is designed to make the agent output ONLY the answer,
-        # without any preambles, thoughts, or "FINAL ANSWER:" tags.
-        # This is critical for exact match scoring.
-        SYSTEM_PROMPT = """You are an ultra-concise and accurate AI assistant for the GAIA benchmark.
-        Your sole purpose is to provide the exact correct answer to the given question, and NOTHING ELSE.
-        Do not include any introductory phrases, thoughts, explanations, or "FINAL ANSWER:" tags.
-        Your output must be the answer only.
-
-        **Here's how you should determine and format your answer:**
-
-        1.  **Understand the Question:** Carefully read the question to identify the precise information needed.
-        2.  **Gather Information:**
-            * Use your internal knowledge if the answer is straightforward or involves simple arithmetic.
-            * **Crucially, use the `DuckDuckGoSearchTool`** if the question requires external knowledge, current facts, or complex research. Formulate the most effective search queries to find exact information.
-            * Analyze search results meticulously to extract the precise fact(s).
-        3.  **Formulate the Answer (Strict Formatting Rules):**
-            * **Numbers:** If the answer is a number, provide it in digits ONLY. Do NOT use commas for thousands separators, currency symbols (e.g., "$"), percentage signs (e.g., "%"), or any other units (e.g., "kg", "meters").
-                * *Example 1 (number):* If the question asks "What is the square root of 81?", your answer should be "9".
-                * *Example 2 (large number):* If the question asks "What is the population of City X (approx)?", and you find "1,234,567", your answer must be "1234567".
-            * **Strings (words):** If the answer is text, use the fewest words possible. Do NOT include articles (a, an, the). Do NOT use abbreviations. If a number is part of a string (e.g., "twenty books"), write the digit in plain English words (e.g., "twenty").
-                * *Example 1 (string):* If the question asks "Who invented the telephone?", your answer should be "Alexander Graham Bell".
-                * *Example 2 (string with number in text):* If the question asks "How many primary colors are there?", your answer should be "three".
-            * **Comma-separated lists:** If the answer is a list of items, apply the above rules to each item, and separate them with a comma and a single space (ee.g., "red, green, blue").
-                * *Example:* If the question asks "List two types of fruit.", your answer should be "apple, banana".
-
-        Your final output *must* be only the answer, following these exact rules. No surrounding text.
+        self.agent = CodeAgent(tools=[DuckDuckGoSearchTool()], model=HfApiModel())
+
+        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.
         """
         self.agent.prompt_templates["system_prompt"] = self.agent.prompt_templates["system_prompt"] + SYSTEM_PROMPT
-            
+        
     def __call__(self, question: str) -> str:
         print(f"Agent received question (first 50 chars): {question[:50]}...")
-        # The key change: The agent's run method should directly produce the final answer,
-        # without additional wrapping, due to the hyper-specific prompt.
         final_answer = self.agent.run(question)
         print(f"Agent returning final answer: {final_answer}")
         return final_answer
@@ -69,6 +41,7 @@ def run_and_submit_all( profile: gr.OAuthProfile | None):
     """
     # --- Determine HF Space Runtime URL and Repo URL ---
     space_id = os.getenv("SPACE_ID") # Get the SPACE_ID for sending link to the code
+
     if profile:
         username= f"{profile.username}"
         print(f"User logged in: {username}")
@@ -86,12 +59,9 @@ def run_and_submit_all( profile: gr.OAuthProfile | None):
     except Exception as e:
         print(f"Error instantiating agent: {e}")
         return f"Error initializing agent: {e}", None
-
     # In the case of an app running as a hugging Face space, this link points toward your codebase ( usefull for others so please keep it public)
-    # IMPORTANT: Update this to YOUR SPACE_ID if you've cloned it!
-    # Ensure this link is correct for your public space for verification.
-    agent_code = f"https://huggingface.co/spaces/{os.getenv('SPACE_ID', 'YOUR_HF_USERNAME/YOUR_SPACE_NAME')}/tree/main"
-    print(f"Agent code will link to: {agent_code}")
+    agent_code = f"https://huggingface.co/spaces/{space_id}/tree/main"
+    print(agent_code)
 
     # 2. Fetch Questions
     print(f"Fetching questions from: {questions_url}")
@@ -126,16 +96,6 @@ def run_and_submit_all( profile: gr.OAuthProfile | None):
             continue
         try:
             submitted_answer = agent(question_text)
-            # IMPORTANT: Post-process the submitted_answer if the model still adds unwanted text.
-            # While the prompt aims to prevent it, sometimes models are stubborn.
-            # This stripping makes sure only the raw answer is sent.
-            submitted_answer = submitted_answer.strip()
-            # Remove any trailing "FINAL ANSWER:" or "FINAL ANSWER:" if model still generates it.
-            if submitted_answer.startswith("FINAL ANSWER:"):
-                submitted_answer = submitted_answer[len("FINAL ANSWER:"):].strip()
-            # You might need more sophisticated stripping depending on how the model misbehaves
-            # For example, if it adds thoughts, you'd need to extract the last line or a specific pattern.
-
             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:
@@ -146,7 +106,7 @@ def run_and_submit_all( profile: gr.OAuthProfile | None):
         print("Agent did not produce any answers to submit.")
         return "Agent did not produce any answers to submit.", pd.DataFrame(results_log)
 
-    # 4. Prepare Submission
+    # 4. Prepare Submission 
     submission_data = {"username": username.strip(), "agent_code": agent_code, "answers": answers_payload}
     status_update = f"Agent finished. Submitting {len(answers_payload)} answers for user '{username}'..."
     print(status_update)
@@ -194,6 +154,7 @@ def run_and_submit_all( profile: gr.OAuthProfile | None):
         results_df = pd.DataFrame(results_log)
         return status_message, results_df
 
+
 # --- Build Gradio Interface using Blocks ---
 with gr.Blocks() as demo:
     gr.Markdown("# Basic Agent Evaluation Runner")
@@ -202,11 +163,15 @@ with gr.Blocks() as demo:
         "Log in to your Hugging Face account using the button below. This uses your HF username for submission. "
         "Click 'Run Evaluation & Submit All Answers' to fetch questions, run your agent, submit answers, and see the score."
     )
+
     gr.LoginButton()
+
     run_button = gr.Button("Run Evaluation & Submit All Answers")
+
     status_output = gr.Textbox(label="Run Status / Submission Result", lines=5, interactive=False)
+    # Removed max_rows=10 from DataFrame constructor
     results_table = gr.DataFrame(label="Questions and Agent Answers", wrap=True)
-    
+
     run_button.click(
         fn=run_and_submit_all,
         outputs=[status_output, results_table]
@@ -214,19 +179,24 @@ with gr.Blocks() as demo:
 
 if __name__ == "__main__":
     print("\n" + "-"*30 + " App Starting " + "-"*30)
+    # Check for SPACE_HOST and SPACE_ID at startup for information
     space_host_startup = os.getenv("SPACE_HOST")
-    space_id_startup = os.getenv("SPACE_ID")
+    space_id_startup = os.getenv("SPACE_ID") # Get SPACE_ID at startup
+
     if space_host_startup:
         print(f"✅ SPACE_HOST found: {space_host_startup}")
         print(f"   Runtime URL should be: https://{space_host_startup}.hf.space")
     else:
         print("ℹ️  SPACE_HOST environment variable not found (running locally?).")
-    if space_id_startup:
+
+    if space_id_startup: # Print repo URLs if SPACE_ID is found
         print(f"✅ SPACE_ID found: {space_id_startup}")
         print(f"   Repo URL: https://huggingface.co/spaces/{space_id_startup}")
         print(f"   Repo Tree URL: https://huggingface.co/spaces/{space_id_startup}/tree/main")
     else:
         print("ℹ️  SPACE_ID environment variable not found (running locally?). Repo URL cannot be determined.")
+
     print("-"*(60 + len(" App Starting ")) + "\n")
+
     print("Launching Gradio Interface for Basic Agent Evaluation...")
     demo.launch(debug=True, share=False)