# backend/news_fetcher.py
from datetime import datetime
import feedparser

def fetch_news(rss_url):
    feed = feedparser.parse(rss_url)
    articles = []
    for entry in feed.entries:
        try:
            # Try parsing with timezone first
            pub_date = datetime.strptime(entry.published, "%a, %d %b %Y %H:%M:%S %z")
        except ValueError:
            try:
                # Fallback to GMT format without timezone
                pub_date = datetime.strptime(entry.published, "%a, %d %b %Y %H:%M:%S %Z")
            except ValueError:
                # Final fallback - use current time if parsing fails
                pub_date = datetime.now()
        
        articles.append({
            "title": entry.title,
            "content": entry.description,
            "published": pub_date,
            "source": rss_url
        })
    return articles