app.py

from smolagents import CodeAgent,DuckDuckGoSearchTool, HfApiModel,load_tool,tool
import datetime
import requests
import pytz
import yaml
from tools.final_answer import FinalAnswerTool
import os
from tavily import TavilyClient
from Gradio_UI import GradioUI

# Below is an example of a tool that does nothing. Amaze us with your creativity !
@tool
def my_custom_tool(arg1:str, arg2:int)-> str: #it's import to specify the return type
    #Keep this format for the description / args / args description but feel free to modify the tool
    """A tool that does nothing yet 
    Args:
        arg1: the first argument
        arg2: the second argument
    """
    return "What magic will you build ?"



# Tavily safe search tool 
@tool
def tavily_search(query: str) -> str:
    """An enhanced search tool that performs web search using Tavily API with better error handling and formatting.
    Args:
        query: The search query to look up
    Returns:
        str: A well-formatted summary of search results with key highlights
    """
    try:
        client = TavilyClient(api_key=os.environ["TAVILY_API_KEY"])
        search_result = client.search(
            query=query, 
            search_depth="advanced",
            include_domains=["*.gov", "*.edu", "*.org"],  # More relaible sources
            exclude_domains=["pinterest.com", "reddit.com"]  # Social media excluded
        )
        
        summary = f"🔍 Search Results for: {query}\n\n"
        summary += "Key Findings:\n"
        
        for idx, result in enumerate(search_result['results'][:5], 1):
            summary += f"\n{idx}. {result['title']}\n"
            summary += f"   Source: {result['domain']}\n"
            if 'published_date' in result:
                summary += f"   Date: {result['published_date']}\n"
            summary += f"   Summary: {result['description'][:250]}...\n"
            
        return summary
    except Exception as e:
        return f"Search Error: {str(e)}. Please try rephrasing your query or try again later."



# Improved time tool
@tool
def world_time_tool(location: str) -> str:
    """An improved timezone tool with better location handling and formatted output.
    Args:
        location: City name or timezone identifier (e.g., 'Tokyo' or 'Asia/Tokyo')
    Returns:
        str: Formatted time information with additional context
    """
    try:
        # Timezone mapping for common cities
        city_to_timezone = {
            'tokyo': 'Asia/Tokyo',
            'new york': 'America/New_York',
            'london': 'Europe/London',
            'paris': 'Europe/Paris',
            'istanbul': 'Europe/Istanbul'
        }
        
        # Clean and check input
        clean_location = location.lower().strip()
        if clean_location in city_to_timezone:
            timezone = city_to_timezone[clean_location]
        else:
            timezone = location
            
        tz = pytz.timezone(timezone)
        current_time = datetime.datetime.now(tz)
        
        # Zengin formatlama
        formatted_time = current_time.strftime("%Y-%m-%d %H:%M:%S")
        day_name = current_time.strftime("%A")
        
        return f"""
            📍 Location: {location.title()}
            🕒 Current Time: {formatted_time}
            📅 Day: {day_name}
        """
    except pytz.exceptions.UnknownTimeZoneError:
        return f"Error: '{location}' is not a recognized timezone or city. Please try a major city name or standard timezone format."
    except Exception as e:
        return f"Time lookup error: {str(e)}"


final_answer = FinalAnswerTool()

# Import tool from Hub
image_generation_tool = load_tool("agents-course/text-to-image", trust_remote_code=True)


# If the agent does not answer, the model is overloaded, please use another model or the following Hugging Face Endpoint that also contains qwen2.5 coder:
model_id='https://pflgm2locj2t89co.us-east-1.aws.endpoints.huggingface.cloud' 

# Model configuration
model = HfApiModel(
    max_tokens=2096,
    temperature=0.3,  # Lower temperature for more consistent responses
    model_id=model_id, # 'Qwen/Qwen2.5-Coder-32B-Instruct',
    custom_role_conversions=None,
)




with open("prompts.yaml", 'r') as stream:
    prompt_templates = yaml.safe_load(stream)
    
agent = CodeAgent(
    model=model,
    tools=[final_answer, tavily_search, get_current_time_in_timezone, image_generation_tool], # Added tavily_search
    max_steps=6,
    verbosity_level=1,
    grammar=None,
    planning_interval=None,
    name=None,
    description=None,
    prompt_templates=prompt_templates
)

# Launch the Gradio interface
GradioUI(agent).launch()