Update app.py

app.py

@@ -1,34 +1,107 @@
 import os
 import gradio as gr
 import requests
-import inspect
 import pandas as pd
+from langchain_community.llms import HuggingFaceHub
+# from dotenv import load_dotenv # Uncomment for local testing with a .env file
+
+# For local testing, you might want to load environment variables from a .env file
+# (ensure .env is in .gitignore and HUGGINGFACEHUB_API_TOKEN is defined in it)
+# if os.path.exists(".env"):
+#     load_dotenv()
 
-# (Keep Constants as is)
 # --- Constants ---
 DEFAULT_API_URL = "https://agents-course-unit4-scoring.hf.space"
 
-# --- Basic Agent Definition ---
+# --- Basic Agent Definition --
 # ----- THIS IS WERE YOU CAN BUILD WHAT YOU WANT ------
 class BasicAgent:
-    def __init__(self):
-        print("BasicAgent initialized.")
+    def __init__(self, hf_api_token: str | None = None):
+        print("BasicAgent initializing...")
+        # Determine the Hugging Face API token
+        # Priority: 1. hf_api_token argument (if passed), 
+        #           2. HUGGINGFACEHUB_API_TOKEN env var, 
+        #           3. HF_TOKEN env var (common for HF Spaces)
+        token_to_use = hf_api_token
+        if not token_to_use:
+            token_to_use = os.getenv("HUGGINGFACEHUB_API_TOKEN")
+        if not token_to_use:
+            token_to_use = os.getenv("HF_TOKEN")
+
+        if not token_to_use:
+            # This error will be caught by the agent instantiation try-except block
+            # in run_and_submit_all, and a message will be shown in the UI.
+            raise ValueError(
+                "Hugging Face API token not found. Please set HUGGINGFACEHUB_API_TOKEN or HF_TOKEN "
+                "as a secret in your Hugging Face Space. This token is required for the LLM."
+            )
+        
+        # You can change the repo_id to any model on the Hugging Face Hub.
+        # Ensure the chosen model is suitable for instruction following / question answering.
+        # Examples: "mistralai/Mistral-7B-Instruct-v0.1", "google/flan-t5-large", "HuggingFaceH4/zephyr-7b-beta"
+        # Using a smaller, faster model for demonstration:
+        self.llm_repo_id = "mistralai/Mistral-7B-Instruct-v0.1"
+        try:
+            self.llm = HuggingFaceHub(
+                repo_id=self.llm_repo_id,
+                model_kwargs={"temperature": 0.1, "max_new_tokens": 150}, # Adjust max_new_tokens as needed
+                huggingfacehub_api_token=token_to_use
+            )
+            print(f"BasicAgent initialized with LLM: {self.llm_repo_id}")
+        except Exception as e:
+            print(f"Error initializing HuggingFaceHub: {e}")
+            raise ValueError(f"Failed to initialize LLM: {e}. Check token and model repo_id.")
+
+
     def __call__(self, question: str) -> str:
-        print(f"Agent received question (first 50 chars): {question[:50]}...")
-        fixed_answer = "This is a default answer."
-        print(f"Agent returning fixed answer: {fixed_answer}")
-        return fixed_answer
+        print(f"Agent received question (first 80 chars): {question[:80]}...")
+
+        # Prompt engineering is crucial.
+        # Instruct the LLM to provide a concise answer without any extra phrases.
+        # Per GAIA instructions: "make sure you don’t include the text “FINAL ANSWER” 
+        # in your submission, just make your agent reply with the answer and nothing else"
+        prompt = f"""Answer the following question concisely. Provide only the answer, without any additional explanation, introductory phrases, or labels like "Answer:".
+
+Question: {question}
+
+Concise Answer:"""
+
+        try:
+            response = self.llm.invoke(prompt)
+            answer = response.strip()
+            
+            # Further cleaning if the model still adds prefixes (common with some models)
+            # Convert to lower for case-insensitive prefix checking
+            answer_lower = answer.lower()
+            common_prefixes = ["answer:", "the answer is:", "concise answer:"]
+            for prefix in common_prefixes:
+                if answer_lower.startswith(prefix):
+                    answer = answer[len(prefix):].strip()
+                    break # Remove only the first matching prefix
+            
+            print(f"Agent LLM raw response (first 80 chars): {response[:80]}...")
+            print(f"Agent final answer (first 80 chars): {answer[:80]}...")
+            
+            if not answer: # Handle cases where the answer becomes empty after stripping
+                print("Warning: Agent produced an empty answer after cleaning.")
+                # Return a placeholder that indicates an issue but is still a string
+                return "Unable to generate a valid answer."
+
+            return answer
+        except Exception as e:
+            print(f"Error during LLM call for question '{question[:50]}...': {e}")
+            # Return an error message string, as the submission expects a string answer.
+            return f"AGENT_ERROR: LLM call failed. ({type(e).__name__})"
 
-def run_and_submit_all( profile: gr.OAuthProfile | None):
+def run_and_submit_all(profile: gr.OAuthProfile | None):
     """
     Fetches all questions, runs the BasicAgent on them, submits all answers,
     and displays the results.
     """
-    # --- Determine HF Space Runtime URL and Repo URL ---
-    space_id = os.getenv("SPACE_ID") # Get the SPACE_ID for sending link to the code
+    space_id = os.getenv("SPACE_ID")
 
     if profile:
-        username= f"{profile.username}"
+        username = f"{profile.username}"
         print(f"User logged in: {username}")
     else:
         print("User not logged in.")
@@ -38,20 +111,22 @@ def run_and_submit_all( profile: gr.OAuthProfile | None):
     questions_url = f"{api_url}/questions"
     submit_url = f"{api_url}/submit"
 
-    # 1. Instantiate Agent ( modify this part to create your agent)
+    # 1. Instantiate Agent
     try:
+        # The BasicAgent will attempt to find the HF token from env variables.
         agent = BasicAgent()
     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)
-    agent_code = f"https://huggingface.co/spaces/{space_id}/tree/main"
-    print(agent_code)
+        # Return the error message to be displayed in the Gradio UI
+        return f"Error initializing agent: {str(e)}", None
+    
+    agent_code = f"https://huggingface.co/spaces/{space_id}/tree/main" if space_id else "local_run_no_space_id"
+    print(f"Agent code link: {agent_code}")
 
     # 2. Fetch Questions
     print(f"Fetching questions from: {questions_url}")
     try:
-        response = requests.get(questions_url, timeout=15)
+        response = requests.get(questions_url, timeout=20) # Increased timeout
         response.raise_for_status()
         questions_data = response.json()
         if not questions_data:
@@ -73,20 +148,26 @@ def run_and_submit_all( profile: gr.OAuthProfile | None):
     results_log = []
     answers_payload = []
     print(f"Running agent on {len(questions_data)} questions...")
-    for item in questions_data:
+    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 task_id or question: {item}")
             continue
+        
+        print(f"\nProcessing question {i+1}/{len(questions_data)}, Task ID: {task_id}")
         try:
             submitted_answer = agent(question_text)
             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:
              print(f"Error running agent on task {task_id}: {e}")
+             # Ensure a placeholder is added for submission to maintain structure
+             error_answer = f"AGENT_RUNTIME_ERROR: {type(e).__name__}"
+             answers_payload.append({"task_id": task_id, "submitted_answer": error_answer})
              results_log.append({"Task ID": task_id, "Question": question_text, "Submitted Answer": f"AGENT ERROR: {e}"})
 
+
     if not answers_payload:
         print("Agent did not produce any answers to submit.")
         return "Agent did not produce any answers to submit.", pd.DataFrame(results_log)
@@ -117,8 +198,8 @@ def run_and_submit_all( profile: gr.OAuthProfile | None):
         try:
             error_json = e.response.json()
             error_detail += f" Detail: {error_json.get('detail', e.response.text)}"
-        except requests.exceptions.JSONDecodeError:
-            error_detail += f" Response: {e.response.text[:500]}"
+        except requests.exceptions.JSONDecodeError: # Renamed from JSONDecodeError for clarity
+            error_detail += f" Response: {e.response.text[:500]}" # Log part of the response
         status_message = f"Submission Failed: {error_detail}"
         print(status_message)
         results_df = pd.DataFrame(results_log)
@@ -133,7 +214,7 @@ def run_and_submit_all( profile: gr.OAuthProfile | None):
         print(status_message)
         results_df = pd.DataFrame(results_log)
         return status_message, results_df
-    except Exception as e:
+    except Exception as e: # Catch any other unexpected errors during submission
         status_message = f"An unexpected error occurred during submission: {e}"
         print(status_message)
         results_df = pd.DataFrame(results_log)
@@ -146,36 +227,56 @@ with gr.Blocks() as demo:
     gr.Markdown(
         """
         **Instructions:**
-
-        1.  Please clone this space, then modify the code to define your agent's logic, the tools, the necessary packages, etc ...
+        1.  This Space uses a `BasicAgent` with an LLM. Ensure you have set your `HUGGINGFACEHUB_API_TOKEN` or `HF_TOKEN` in the Space secrets for the LLM to work.
         2.  Log in to your Hugging Face account using the button below. This uses your HF username for submission.
         3.  Click 'Run Evaluation & Submit All Answers' to fetch questions, run your agent, submit answers, and see the score.
-
         ---
         **Disclaimers:**
-        Once clicking on the "submit button, it can take quite some time ( this is the time for the agent to go through all the questions).
-        This space provides a basic setup and is intentionally sub-optimal to encourage you to develop your own, more robust solution. For instance for the delay process of the submit button, a solution could be to cache the answers and submit in a seperate action or even to answer the questions in async.
+        Once clicking on the "submit" button, it can take quite some time (this is the time for the agent to go through all the questions using an LLM).
+        This space provides a basic setup. For better GAIA scores, you might need to:
+            - Choose a more powerful LLM.
+            - Improve prompt engineering.
+            - Implement tool usage for questions requiring file access or external actions (the API provides `/files/{task_id}`).
         """
     )
 
+    # Session state to hold the Hugging Face profile (token and username)
+    # This isn't strictly necessary for this version as token is read from env for LLM
+    # but good practice if profile info is needed elsewhere.
+    hf_profile_state = gr.State(None) 
+    
+    # Wrap LoginButton with a function to capture the profile
+    def login_handler(profile: gr.OAuthProfile | None):
+        if profile:
+            print(f"Profile captured: {profile.username}")
+            # If you wanted to pass profile.token to agent:
+            # BasicAgent(hf_api_token=profile.token) - but env var method is preferred for LLM token
+        return profile
+
+    # The gr.LoginButton() automatically provides the profile to functions that list it as an input
+    # So, `run_and_submit_all` will receive it directly when triggered by `run_button`.
+    # No explicit state management for profile passing to `run_and_submit_all` is needed here.
     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)
+    results_table = gr.DataFrame(label="Questions and Agent Answers", wrap=True) # Removed max_rows
 
+    # The profile from gr.LoginButton() is implicitly passed as the first argument
+    # to `run_and_submit_all` if its signature includes it.
     run_button.click(
         fn=run_and_submit_all,
+        # No explicit inputs needed here if `gr.LoginButton` handles profile passing.
+        # If explicit passing was needed from a state: inputs=[hf_profile_state],
         outputs=[status_output, results_table]
     )
 
 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") # Get SPACE_ID at startup
+    space_id_startup = os.getenv("SPACE_ID")
 
     if space_host_startup:
         print(f"✅ SPACE_HOST found: {space_host_startup}")
@@ -183,12 +284,18 @@ if __name__ == "__main__":
     else:
         print("ℹ️  SPACE_HOST environment variable not found (running locally?).")
 
-    if space_id_startup: # Print repo URLs if SPACE_ID is found
+    if space_id_startup:
         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.")
+    
+    # Check for HF_TOKEN at startup as a hint for the user
+    if not (os.getenv("HUGGINGFACEHUB_API_TOKEN") or os.getenv("HF_TOKEN")):
+        print("⚠️  WARNING: HUGGINGFACEHUB_API_TOKEN or HF_TOKEN environment variable not found.")
+        print("   The LLM agent will likely fail to initialize. Please set this token in your Space secrets.")
+
 
     print("-"*(60 + len(" App Starting ")) + "\n")