from langchain_community.document_loaders import WikipediaLoader
from langchain_community.document_loaders import ArxivLoader
from langchain_community.retrievers import WikipediaRetriever
from langchain_community.tools.tavily_search import TavilySearchResults
from langchain_community.tools import DuckDuckGoSearchResults
from langchain_core.tools import tool
import io
import openpyxl
import os
# from smolagents import tool
import requests
from PIL import Image
from bs4 import BeautifulSoup
@tool
def web_search(query: str) -> dict:
"""Search Tavily for a query and return maximum 3 results.
Args:
query: The search query."""
search_docs = DuckDuckGoSearchResults(max_results=3, output_format='list').invoke(query)
formatted_search_docs = "\n\n---\n\n".join(
[
f'\n{doc.get("snippet", "")}\n'
for doc in search_docs
]
)
return {"web_results": formatted_search_docs}
# @tool
# def wiki_search(query: str) -> dict:
# """Search Wikipedia for a query and return maximum 2 results.
# Args:
# query: The search query."""
# search_docs = WikipediaLoader(query=query, load_max_docs=2).load()
# formatted_search_docs = "\n\n---\n\n".join(
# [
# f'\n{doc.page_content}\n'
# for doc in search_docs
# ])
# return {"wiki_results": formatted_search_docs}
# @tool
# def wiki_search(query: str) -> dict:
# """Search Wikipedia for a query and return maximum 2 results.
# Args:
# query: The search query."""
# search_docs = WikipediaRetriever(load_max_docs=5).invoke(query)
# formatted_search_docs = "\n\n---\n\n".join(
# [
# f'\n{doc.page_content}\n'
# for doc in search_docs
# ])
# return {"wiki_results": formatted_search_docs}
@tool
def wiki_search(query: str) -> dict:
"""Search Wikipedia for a query and return maximum 1 results.
Args:
query: The search query."""
search_docs = WikipediaRetriever(load_max_docs=1, top_k_results=2).invoke(query)
wiki_results = []
for doc in search_docs:
url = doc.metadata["source"] if doc else ""
print(url)
response = requests.get(url)
response_text = response.text
soup = BeautifulSoup(response_text, "html.parser")
wiki_results.append(' '.join(soup.get_text().split())[:20000])
return {"wiki_results": wiki_results}
@tool
def reverse_string(query: str) -> dict:
"""Reverse the input string.
Args:
query: The input string to reverse."""
return {"reversed_string": query[::-1]}
@tool
def calculator(expression: str) -> dict:
"""Perform mathematical calculations and return the result.
This calculator can handle:
- Basic arithmetic: +, -, *, /, % (modulus)
- Parentheses for order of operations
- Decimal numbers
- Multiple operations in one expression
Args:
expression: A mathematical expression as a string
Returns:
A string containing the calculation result
Examples:
calculator("25 * 4") -> "100"
calculator("100 / 5") -> "20.0"
calculator("(15 + 30) * 2") -> "90"
calculator("50 - 20 + 10") -> "40"
calculator("17 % 5") -> "2"
calculator("100 % 7") -> "2"
calculator("(20 + 5) % 8") -> "5"
"""
try:
# Clean the expression
expression = expression.strip()
# Validate that the expression only contains safe characters (now including %)
allowed_chars = set('0123456789+-*/.()% ')
if not all(c in allowed_chars for c in expression):
raise ValueError("Expression contains invalid characters. Only numbers and +, -, *, /, %, (, ) are allowed.")
result = eval(expression)
# Format the result
if isinstance(result, float) and result.is_integer():
return str(int(result))
else:
return str(result)
except ZeroDivisionError:
return "Error: Cannot divide by zero or modulus by zero"
except SyntaxError:
return f"Error: Invalid mathematical expression: {expression}"
except Exception as e:
return f"Error: {str(e)}"
@tool
def tool_read_files(filepath: str) -> str:
"""
Downloads a .py or .xlsx file from a remote URL and returns its contents as plain text.
Raises a recoverable exception if the file does not end with .py or .xlsx.
Args:
filepath: The path to the Python (.py) or Excel (.xlsx) file.
"""
root_url = "https://agents-course-unit4-scoring.hf.space/files/"
# Strip the file extension from the url before downloading
base, ext = os.path.splitext(filepath)
url = root_url + base
if filepath.endswith('.py'):
response = requests.get(url)
if response.status_code != 200:
raise Exception(f"Recoverable: Failed to download file from {url}")
return response.text
elif filepath.endswith('.xlsx'):
response = requests.get(url)
if response.status_code != 200:
raise Exception(f"Recoverable: Failed to download file from {url}")
wb = openpyxl.load_workbook(io.BytesIO(response.content), data_only=True)
result = []
for sheet in wb.worksheets:
result.append(f"# Sheet: {sheet.title}")
for row in sheet.iter_rows(values_only=True):
result.append(','.join([str(cell) if cell is not None else '' for cell in row]))
return '\n'.join(result)
else:
raise Exception("Recoverable: Only .py and .xlsx files can be read with this tool.")
@tool
def tool_download_image(filepath: str) -> str:
"""
Downloads an image file (.png, .jpg, .jpeg) from a remote URL and returns useful information about the image.
This includes the image URL and basic metadata like dimensions and format.
Raises a recoverable exception if the file is not a supported image type.
Args:
filepath: The path to the image file.
"""
root_url = "https://agents-course-unit4-scoring.hf.space/files/"
base, ext = os.path.splitext(filepath)
url = root_url + base
if ext.lower() in ['.png', '.jpg', '.jpeg']:
response = requests.get(url)
if response.status_code != 200:
raise Exception(f"Recoverable: Failed to download image from {url}")
# Get image metadata using Pillow
try:
img = Image.open(io.BytesIO(response.content))
width, height = img.size
format = img.format
mode = img.mode
# Return useful information about the image
return f"Image URL: {url}\nFormat: {format}\nDimensions: {width}x{height}\nMode: {mode}"
except ImportError:
# Fallback if PIL is not available
content_type = response.headers.get('Content-Type', 'unknown')
content_length = response.headers.get('Content-Length', 'unknown')
return f"Content-Type: {content_type}\nSize: {content_length} bytes"
else:
raise Exception("Recoverable: Only .png, .jpg, and .jpeg files can be processed with this tool.")