from smolagents import CodeAgent, DuckDuckGoSearchTool, HfApiModel, LiteLLMModel, load_tool, tool
import datetime
import requests
import pytz
import yaml
from tools.final_answer import FinalAnswerTool
import yfinance as yf
from transformers import pipeline
from newspaper import Article
import numpy as np
import os

from Gradio_UI import GradioUI

# Below is an example of a tool that retrieves news for a given company and performs sentiment analysis on the articles!

# Initialize sentiment analysis pipeline
sentiment_analyzer = pipeline("sentiment-analysis", model="distilbert-base-uncased-finetuned-sst-2-english")

def analyze_article(url):
    # Extract article 
    article = Article(url)
    article.download()
    article.parse()
    
    # Process in chunks if article is long (max 512 tokens for BERT-based models)
    chunks = [article.text[i:i+500] for i in range(0, len(article.text), 500)]
    results = sentiment_analyzer(chunks)
    
    # Extract scores and calculate average
    scores = [result['score'] * (1 if result['label'] == 'POSITIVE' else -1) for result in results]
    return {
        "title": article.title,
        "article": article.text[:100],
        "sentiment score": f"{np.mean(scores):.2f}",        
        "summary": f"{'Positive' if np.mean(scores) > 0 else 'Negative'} sentiment detected"
    }

@tool
def fetch_news(company:str, count:int=3)-> list: #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 gets the latest news articles for a given company.
    Args:
        company: name or trading symbol of the company
        count: number of news articles to fetch
    """
    try:
        news = yf.Search(query=company, news_count=count).news

        print(f"Latest news articles with sentiment for {company}:")
        
        # Return list of articles with sentiment detected
        response = [analyze_article(n['link']) for n in news]
        
        return response
    except Exception as e:
        return f"Error fetching news for company '{company}': {str(e)}"

@tool
def get_current_time_in_timezone(timezone: str) -> str:
    """A tool that fetches the current local time in a specified timezone.
    Args:
        timezone: A string representing a valid timezone (e.g., 'America/New_York').
    """
    try:
        # Create timezone object
        tz = pytz.timezone(timezone)
        # Get current time in that timezone
        local_time = datetime.datetime.now(tz).strftime("%Y-%m-%d %H:%M:%S")
        return f"The current local time in {timezone} is: {local_time}"
    except Exception as e:
        return f"Error fetching time for timezone '{timezone}': {str(e)}"


final_answer = FinalAnswerTool()

# 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 = HfApiModel(
max_tokens=2096,
temperature=0.5,
model_id='deepseek-ai/DeepSeek-R1', #'Qwen/Qwen2.5-Coder-32B-Instruct',# it is possible that this model may be overloaded
custom_role_conversions=None,
)
"""
google_api_key= os.getenv("GOOGLE_API_KEY")

model = LiteLLMModel(
max_tokens=100,
temperature=0.5,
model_id='gemini/gemini-1.5-flash', #'Qwen/Qwen2.5-Coder-32B-Instruct',# it is possible that this model may be overloaded
api_key = google_api_key,
custom_role_conversions=None,
)

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

with open("prompts.yaml", 'r') as stream:
    prompt_templates = yaml.safe_load(stream)
    
agent = CodeAgent(
    model=model,
    tools=[final_answer, fetch_news, image_generation_tool, get_current_time_in_timezone], ## add your tools here (don't remove final answer)
    max_steps=6,
    verbosity_level=1,
    grammar=None,
    planning_interval=None,
    name=None,
    description=None,
    prompt_templates=prompt_templates
)


GradioUI(agent).launch()