Update app.py

app.py

@@ -3,68 +3,226 @@ import gradio as gr
 import requests
 import inspect
 import pandas as pd
-
-from smolagents import CodeAgent, InferenceClientModel, WebSearchTool
-
+import tempfile
+ 
+from smolagents import CodeAgent, InferenceClientModel, WebSearchTool, tool
+ 
 # (Keep Constants as is)
 # --- Constants ---
 DEFAULT_API_URL = "https://agents-course-unit4-scoring.hf.space"
-
+ 
+ 
+# --- Custom tools for reading task attachments ---
+ 
+def _download_task_file(task_id: str) -> str:
+    """Internal helper: downloads the file attached to a task_id and saves
+    it to a temp folder, returning the local file path (or '' if none)."""
+    url = f"{DEFAULT_API_URL}/files/{task_id}"
+    try:
+        response = requests.get(url, timeout=30)
+        if response.status_code != 200:
+            return ""
+        # Try to get a filename from the Content-Disposition header
+        cd = response.headers.get("content-disposition", "")
+        filename = task_id
+        if "filename=" in cd:
+            filename = cd.split("filename=")[-1].strip('"; ')
+        else:
+            # Guess an extension from content-type
+            ctype = response.headers.get("content-type", "")
+            if "spreadsheet" in ctype or "excel" in ctype:
+                filename = f"{task_id}.xlsx"
+            elif "audio" in ctype:
+                filename = f"{task_id}.mp3"
+            elif "pdf" in ctype:
+                filename = f"{task_id}.pdf"
+            elif "csv" in ctype:
+                filename = f"{task_id}.csv"
+        tmp_dir = tempfile.gettempdir()
+        file_path = os.path.join(tmp_dir, filename)
+        with open(file_path, "wb") as f:
+            f.write(response.content)
+        return file_path
+    except Exception as e:
+        print(f"Error downloading file for task {task_id}: {e}")
+        return ""
+ 
+ 
+@tool
+def download_task_file(task_id: str) -> str:
+    """Downloads the file attached to a benchmark question (identified by its task_id)
+    and saves it locally. Use this FIRST whenever a question mentions an attached file
+    (Excel, CSV, audio, image, PDF, python script, etc).
+ 
+    Args:
+        task_id: The task_id of the current question.
+ 
+    Returns:
+        The local file path where the file was saved, or an empty string if there is
+        no file for this task_id.
+    """
+    return _download_task_file(task_id)
+ 
+ 
+@tool
+def read_excel_file(file_path: str) -> str:
+    """Reads an Excel (.xlsx/.xls) file and returns its content as readable text
+    (one table per sheet). Use this after downloading the file with download_task_file.
+ 
+    Args:
+        file_path: Local path to the Excel file.
+ 
+    Returns:
+        A text representation of every sheet in the workbook.
+    """
+    try:
+        sheets = pd.read_excel(file_path, sheet_name=None, engine="openpyxl")
+        out = []
+        for name, df in sheets.items():
+            out.append(f"--- Sheet: {name} ---\n{df.to_string(index=False)}")
+        return "\n\n".join(out)
+    except Exception as e:
+        return f"Error reading Excel file: {e}"
+ 
+ 
+@tool
+def read_csv_file(file_path: str) -> str:
+    """Reads a CSV or TSV file and returns its content as readable text.
+ 
+    Args:
+        file_path: Local path to the CSV file.
+ 
+    Returns:
+        A text representation of the table.
+    """
+    try:
+        df = pd.read_csv(file_path)
+        return df.to_string(index=False)
+    except Exception as e:
+        return f"Error reading CSV file: {e}"
+ 
+ 
+@tool
+def read_text_file(file_path: str) -> str:
+    """Reads a plain text, code, or markdown file and returns its raw content.
+ 
+    Args:
+        file_path: Local path to the text/code file.
+ 
+    Returns:
+        The raw text content of the file.
+    """
+    try:
+        with open(file_path, "r", encoding="utf-8", errors="replace") as f:
+            return f.read()
+    except Exception as e:
+        return f"Error reading text file: {e}"
+ 
+ 
+@tool
+def transcribe_audio_file(file_path: str) -> str:
+    """Transcribes a speech audio file (mp3/wav) to text using a Whisper model.
+ 
+    Args:
+        file_path: Local path to the audio file.
+ 
+    Returns:
+        The transcribed text content of the audio.
+    """
+    try:
+        from transformers import pipeline
+        asr = pipeline("automatic-speech-recognition", model="openai/whisper-base")
+        result = asr(file_path)
+        return result.get("text", "") if isinstance(result, dict) else str(result)
+    except Exception as e:
+        return f"Error transcribing audio file: {e}"
+ 
+ 
 # --- Basic Agent Definition ---
 # ----- THIS IS WERE YOU CAN BUILD WHAT YOU WANT ------
 class BasicAgent:
     """
     A real agent built with smolagents.
-    Uses a free Hugging Face hosted model + a web search tool + the
-    built-in Python code execution that CodeAgent already provides.
+    Uses a free Hugging Face hosted model, web search, and a set of file
+    tools (Excel/CSV/text/audio) to handle GAIA-style benchmark questions
+    that come with an attachment.
     """
     def __init__(self):
         print("BasicAgent initializing...")
-
+ 
         # Free model hosted by Hugging Face Inference Providers.
         # No paid API key required - just needs the Space's default HF token.
         self.model = InferenceClientModel(
             model_id="Qwen/Qwen2.5-Coder-32B-Instruct",
         )
-
+ 
         self.agent = CodeAgent(
-            tools=[WebSearchTool()],
+            tools=[
+                WebSearchTool(),
+                download_task_file,
+                read_excel_file,
+                read_csv_file,
+                read_text_file,
+                transcribe_audio_file,
+            ],
             model=self.model,
             add_base_tools=True,   # adds python interpreter + a couple of extra default tools
-            max_steps=8,
+            additional_authorized_imports=[
+                "pandas", "numpy", "json", "re", "math", "datetime",
+                "openpyxl", "io", "csv",
+            ],
+            max_steps=10,
         )
-
+ 
         print("BasicAgent initialized.")
-
-    def __call__(self, question: str) -> str:
+ 
+    def __call__(self, question: str, task_id: str = "") -> str:
         print(f"Agent received question (first 50 chars): {question[:50]}...")
-
+ 
+        # Some GAIA questions are written backwards as a "riddle" test.
+        # Detect this and flip it back before sending to the model.
+        reversed_hint = ""
+        if question.strip().endswith(".") and question.strip()[:1].islower():
+            # crude heuristic: try reversing and see if it reads like English
+            flipped = question.strip()[::-1]
+            if flipped[:1].isupper() or flipped.split(" ")[0].isalpha():
+                reversed_hint = (
+                    f"\n\nNote: this question may be written backwards. "
+                    f"Reversed, it reads: {flipped}"
+                )
+ 
         # Strong instruction to keep answers in the exact-match format
         # the GAIA benchmark expects: no "FINAL ANSWER" prefix, no extra
         # explanation, just the bare answer.
         instructions = (
             "You are a general AI assistant answering a benchmark question. "
+            f"The task_id for this question is '{task_id}'. If the question "
+            "mentions an attached file (Excel, CSV, audio, image, code, etc.), "
+            "call download_task_file('" + task_id + "') first to get its local "
+            "path, then use the matching reading tool (read_excel_file, "
+            "read_csv_file, read_text_file, or transcribe_audio_file) on that path.\n\n"
             "Report your thoughts, then finish with the answer. "
             "Your final output must be ONLY the answer itself: "
             "no explanations, no extra words, no 'FINAL ANSWER' prefix. "
             "If the answer is a number, write only the number (no units unless "
             "explicitly requested). If it's a string, give the minimal exact phrase "
             "requested, avoiding articles and abbreviations unless asked otherwise. "
-            "If it's a list, give a comma separated list following the same rules.\n\n"
+            "If it's a list, give a comma separated list following the same rules."
+            f"{reversed_hint}\n\n"
             f"Question: {question}"
         )
-
+ 
         try:
             result = self.agent.run(instructions)
             answer = str(result).strip()
         except Exception as e:
             print(f"Agent error while answering: {e}")
             answer = "I don't know."
-
+ 
         print(f"Agent returning answer: {answer}")
         return answer
-
-
+ 
+ 
 def run_and_submit_all( profile: gr.OAuthProfile | None):
     """
     Fetches all questions, runs the BasicAgent on them, submits all answers,
@@ -72,18 +230,18 @@ 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}")
     else:
         print("User not logged in.")
         return "Please Login to Hugging Face with the button.", None
-
+ 
     api_url = DEFAULT_API_URL
     questions_url = f"{api_url}/questions"
     submit_url = f"{api_url}/submit"
-
+ 
     # 1. Instantiate Agent ( modify this part to create your agent)
     try:
         agent = BasicAgent()
@@ -93,7 +251,7 @@ def run_and_submit_all( profile: gr.OAuthProfile | 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)
-
+ 
     # 2. Fetch Questions
     print(f"Fetching questions from: {questions_url}")
     try:
@@ -114,7 +272,7 @@ def run_and_submit_all( profile: gr.OAuthProfile | None):
     except Exception as e:
         print(f"An unexpected error occurred fetching questions: {e}")
         return f"An unexpected error occurred fetching questions: {e}", None
-
+ 
     # 3. Run your Agent
     results_log = []
     answers_payload = []
@@ -126,22 +284,22 @@ def run_and_submit_all( profile: gr.OAuthProfile | None):
             print(f"Skipping item with missing task_id or question: {item}")
             continue
         try:
-            submitted_answer = agent(question_text)
+            submitted_answer = agent(question_text, task_id=task_id)
             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}")
              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)
-
+ 
     # 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)
-
+ 
     # 5. Submit
     print(f"Submitting {len(answers_payload)} answers to: {submit_url}")
     try:
@@ -184,59 +342,59 @@ def run_and_submit_all( profile: gr.OAuthProfile | None):
         print(status_message)
         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")
     gr.Markdown(
         """
         **Instructions:**
-
+ 
         1.  Please clone this space, then modify the code to define your agent's logic, the tools, the necessary packages, etc ...
         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.
         """
     )
-
+ 
     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]
     )
-
+ 
 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
-
+ 
     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: # 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, ssr_mode=False)
+    demo.launch(debug=True, share=False, ssr_mode=False)