feat: added Excel Tool

agent.py

@@ -1,37 +1,104 @@
 import os
+from pathlib import Path
+from typing import Optional, Union
 
+import pandas as pd
 from dotenv import load_dotenv
-from smolagents import (CodeAgent, DuckDuckGoSearchTool, FinalAnswerTool, LiteLLMModel,
-                        PythonInterpreterTool, VisitWebpageTool, WikipediaSearchTool)
+from smolagents import (
+    CodeAgent,
+    DuckDuckGoSearchTool,
+    FinalAnswerTool,
+    LiteLLMModel,
+    PythonInterpreterTool,
+    WikipediaSearchTool,
+)
+from smolagents.tools import Tool
+from tabulate import tabulate
 
+# Load environment variables
 load_dotenv()
 
+# Initialize the model
 model = LiteLLMModel(
     model_id=os.getenv("GEMINI_MODEL"),
     api_key=os.getenv("GEMINI_API_KEY")
+)
+
+
+class ExcelToTextTool(Tool):
+    """Render an Excel worksheet as a Markdown table."""
+
+    name = "excel_to_text"
+    description = (
+        "Read an Excel file and return a Markdown table of the requested sheet. "
+        "Accepts either the sheet name or a zero-based index (as a string)."
     )
 
+    inputs = {
+        "excel_path": {
+            "type": "string",
+            "description": "Path to the Excel file (.xlsx or .xls).",
+        },
+        "sheet_name": {
+            "type": "string",
+            "description": (
+                "Worksheet name or zero-based index (as a string). "
+                "Optional; defaults to the first sheet."
+            ),
+            "nullable": True,
+        },
+    }
+
+    output_type = "string"
+
+    def forward(self, excel_path: str, sheet_name: Optional[str] = None) -> str:
+        """Load the Excel file and return the sheet as a Markdown table."""
+
+        file_path = Path(excel_path).expanduser().resolve()
+        if not file_path.is_file():
+            return f"Error: Excel file not found at {file_path}"
+
+        try:
+            sheet: Union[str, int] = (
+                int(
+                    sheet_name
+                    ) if sheet_name and sheet_name.isdigit() else sheet_name or 0
+            )
+
+            df = pd.read_excel(file_path, sheet_name=sheet)
+
+            if hasattr(df, "to_markdown"):
+                return df.to_markdown(index=False)
+
+            return tabulate(df, headers="keys", tablefmt="github", showindex=False)
+
+        except Exception as e:
+            return f"Error reading Excel file: {e}"
+
 
 class GaiaAgent:
-    """
-    An agent designed to answer questions using a combination of tools,
-    including search engines, web page access, a Python interpreter, and more.
-    """
+    """An agent capable of using tools to answer general questions."""
+
     def __init__(self):
         print("GaiaAgent initialized with tools.")
 
         tools = [
             DuckDuckGoSearchTool(),
-            VisitWebpageTool(),
             WikipediaSearchTool(),
+            ExcelToTextTool(),
             PythonInterpreterTool(),
             FinalAnswerTool(),
         ]
 
-        self.agent = CodeAgent(model=model, tools=tools)
+        self.agent = CodeAgent(
+            model=model,
+            tools=tools,
+            add_base_tools=True,
+            additional_authorized_imports=["pandas", "numpy", "csv", "subprocess"],
+        )
 
     def __call__(self, task_id: str, question: str) -> str:
-        print(f"Agent received {task_id=}\n{question[:50]=}...")
+        print(f"Agent received task_id='{task_id}' | question='{question[:50]}...'")
         answer = self.agent.run(question)
         print(f"Agent returning answer: {answer}")
         return answer

logic.py

@@ -1,5 +1,7 @@
 from typing import Dict, List, Tuple
-
+import re
+import tempfile
+from pathlib import Path
 import pandas as pd
 import requests
 from agent import GaiaAgent
@@ -9,6 +11,7 @@ from pandas import DataFrame
 DEFAULT_API_URL = "https://agents-course-unit4-scoring.hf.space"
 QUESTIONS_URL = f"{DEFAULT_API_URL}/questions"
 SUBMIT_URL = f"{DEFAULT_API_URL}/submit"
+FILE_PATH = f"{DEFAULT_API_URL}/files/"
 
 
 # --- Helper Methods ---
@@ -142,6 +145,7 @@ def run_agent(
     for item in questions_data:
         task_id = item.get("task_id")
         question_text = item.get("question")
+        question_text = process_file(task_id, question_text)
         if not task_id or question_text is None:
             print(f"⚠️ Skipping invalid item (missing task_id or question): {item}")
             continue
@@ -162,3 +166,42 @@ def run_agent(
             }
         )
     return results_log, answers_payload
+
+
+def process_file(task_id: str, question_text: str) -> str:
+    """
+    Attempt to download a file associated with a task from the API.
+
+    - If the file exists (HTTP 200), it is saved to a temp directory and the local file path is returned.
+    - If no file is found (HTTP 404), returns None.
+    - For all other HTTP errors, the exception is propagated to the caller.
+    """
+    file_url = f"{FILE_PATH}{task_id}"
+
+    try:
+        response = requests.get(file_url, timeout=30)
+        response.raise_for_status()
+    except requests.exceptions.RequestException as exc:
+        print(f"Exception in download_file>> {str(exc)}")
+        return question_text # Unable to get the file
+
+    # Determine filename from 'Content-Disposition' header, fallback to task_id
+    content_disposition = response.headers.get("content-disposition", "")
+    filename = task_id
+    match = re.search(r'filename="([^"]+)"', content_disposition)
+    if match:
+        filename = match.group(1)
+
+    # Save file in a temp directory
+    temp_storage_dir = Path(tempfile.gettempdir()) / "gaia_cached_files"
+    temp_storage_dir.mkdir(parents=True, exist_ok=True)
+
+    file_path = temp_storage_dir / filename
+    file_path.write_bytes(response.content)
+    return (
+                f"{question_text}\n\n"
+                f"---\n"
+                f"A file was downloaded for this task and saved locally at:\n"
+                f"{str(file_path)}\n"
+                f"---\n\n"
+            )

requirements.txt

@@ -6,4 +6,5 @@ pandas
 smolagents
 wikipedia-api
 google-generativeai
-smolagents[litellm]
+smolagents[litellm]
+tabulate