add tool to read file and run python code

Agents/agent.py

@@ -19,7 +19,7 @@ from langchain_core.messages import (
     convert_to_messages,
 )
 from pydantic import BaseModel
-from Tools.tools import webSearch, youtubeVideoTranscript, arxivSearch, wikiSearch,add,multiply,divide,substract,modulus,power,count_substring
+from Tools.tools import webSearch, youtubeVideoTranscript, arxivSearch, wikiSearch,add,multiply,divide,substract,modulus,power,count_substring,run_python_code_from_url,read_excel_from_url,transcribe_audio,read_image_file
 from langchain_core.messages import SystemMessage, HumanMessage
 from dotenv import load_dotenv
 from supabase.client import Client, create_client
@@ -127,7 +127,11 @@ tools = [
     modulus,
     power,
     count_substring,
-    youtubeVideoTranscript
+    youtubeVideoTranscript,
+    run_python_code_from_url,
+    read_excel_from_url,
+    transcribe_audio,
+    read_image_file,
 ]
 
 def tools_condition1(

Tools/tools.py

@@ -8,6 +8,17 @@ from langchain.tools.retriever import create_retriever_tool
 from langchain_community.document_loaders import YoutubeLoader
 from langchain_community.document_loaders.youtube import TranscriptFormat
 from langchain_tavily import TavilySearch
+import base64
+from groq import Groq
+import os
+import pandas as pd
+import requests
+from io import BytesIO
+import sys
+import io
+import traceback
+
+gaiaValidationURL = os.getenv("GaiaValidationURL")
 
 @tool
 def add(a: int, b:int) -> int:
@@ -152,5 +163,125 @@ def count_substring(substring:str, text:str) -> int:
     """
     return text.count(substring)
 
+@tool
+def read_image_file(file_path: str) -> dict:
+    """
+    Reads an image file and returns a dict containing base64-encoded image data.
+    This can be passed to a Groq vision-enabled LLM message.
+
+    Args:
+        file_path (str): Path to the image file.
+
+    Returns:
+        dict: {"type": "image_url", "image_url": {"url": "data:image/<ext>;base64,<data>"}}
+    """
+    try:
+        formattedFilePath = gaiaValidationURL+ file_path.lower()
+        print(f"excel url {formattedFilePath}")
+
+        ext = formattedFilePath.lower().lstrip(".")
+        if ext not in ["jpg", "jpeg", "png", "webp"]:
+            return {"error": f"Unsupported image format: {ext}"}
+
+        with open(formattedFilePath, "rb") as f:
+            b64_data = base64.b64encode(f.read()).decode("utf-8")
+
+        return {
+            "type": "image_url",
+            "image_url": {
+                "url": f"data:image/{ext};base64,{b64_data}"
+            }
+        }
+    except Exception as e:
+        return {"error": str(e)}
+    
+@tool
+def transcribe_audio(file_path: str) -> str:
+    """
+    Transcribes an audio file (mp3, wav, m4a, etc.) using Groq Whisper model.
+    
+    Args:
+        file_path (str): Path to the audio file.
+        
+    Returns:
+        str: Transcribed text from the audio.
+    """
+    try:
+        formattedFilePath = gaiaValidationURL+ file_path.lower()
+        print(f"excel url {formattedFilePath}")
+
+        ext = formattedFilePath.lower().lstrip(".")
+        if ext not in ["mp3", "wav", "m4a"]:
+            return {"error": f"Unsupported image format: {ext}"}
+        
+        client =  Groq(api_key=os.getenv("GROQ_API_KEY"))
+        
+        with open(formattedFilePath, "rb") as f:
+            transcription = client.audio.transcriptions.create(
+                model="whisper-large-v3",
+                file=f
+            )
+        
+        return transcription.text
+    except Exception as e:
+        return f"Error transcribing audio: {str(e)}"
+    
+@tool
+def read_excel_from_url(url: str, as_json: bool = False) -> str:
+    """
+    Downloads an Excel file from a URL and returns its contents as text or JSON.
+
+    Args:
+        url (str): URL to the Excel file.
+        as_json (bool): If True, return JSON string; otherwise plain text.
+
+    Returns:
+        str: Contents of the Excel file as text or JSON.
+    """
+    try:
+        response = requests.get(gaiaValidationURL + url)
+        print(f"excel url {gaiaValidationURL + url}")
+        response.raise_for_status()
+        data = BytesIO(response.content)
+        df = pd.read_excel(data)
+
+        if as_json:
+            return df.to_json(orient="records")
+        else:
+            return df.to_string(index=False)
+    except Exception as e:
+        return f"Error reading Excel file from URL: {str(e)}"
+    
+@tool
+def run_python_code_from_url(url: str) -> str:
+    """
+    Downloads Python code from a URL, executes it, and returns the output or errors.
+
+    Args:
+        url (str): URL of the Python code to execute.
+
+    Returns:
+        str: Captured output or error traceback.
+    """
+    try:
+        formattedURL = gaiaValidationURL + url
+        response = requests.get(formattedURL)
+        print(f"pythonurl : {formattedURL}")
+        response.raise_for_status()
+        code = response.text
+
+        old_stdout = sys.stdout
+        sys.stdout = mystdout = io.StringIO()
+
+        exec_globals = {}
+        exec(code, exec_globals)
+
+        sys.stdout = old_stdout
 
-    
+        output = mystdout.getvalue()
+        if not output.strip():
+            output = "Code executed successfully with no output."
+        return output
+    except Exception:
+        sys.stdout = old_stdout
+        return "Error executing code:\n" + traceback.format_exc()

requirements.txt

@@ -21,4 +21,5 @@ pytesseract
 matplotlib
 langfuse
 typing
-pydantic
+pydantic
+groq