flights_server.py

from mcp.server.fastmcp import FastMCP
from datetime import datetime
import random
import urllib.parse

# Initialize FastMCP server for Flights
mcp = FastMCP("FlightsAgent")

# Real airline data for realistic results
REAL_AIRLINES = [
    {"name": "Emirates", "code": "EK", "hub": "DXB"},
    {"name": "British Airways", "code": "BA", "hub": "LHR"},
    {"name": "Qatar Airways", "code": "QR", "hub": "DOH"},
    {"name": "Turkish Airlines", "code": "TK", "hub": "IST"},
    {"name": "Lufthansa", "code": "LH", "hub": "FRA"},
    {"name": "Air France", "code": "AF", "hub": "CDG"},
    {"name": "KLM", "code": "KL", "hub": "AMS"},
    {"name": "Etihad Airways", "code": "EY", "hub": "AUH"},
    {"name": "Swiss", "code": "LX", "hub": "ZRH"},
    {"name": "Singapore Airlines", "code": "SQ", "hub": "SIN"},
]

# Common layover cities
LAYOVER_CITIES = {
    "DOH": "Doha", "IST": "Istanbul", "FRA": "Frankfurt", 
    "CDG": "Paris", "AMS": "Amsterdam", "DXB": "Dubai",
    "AUH": "Abu Dhabi", "ZRH": "Zurich", "MUC": "Munich"
}

def generate_realistic_price(origin: str, destination: str, passengers: int, cabin_class: str = "economy") -> int:
    """Generate realistic flight prices based on route."""
    # Base prices for common routes (economy, per person)
    base_prices = {
        "short": (150, 350),      # < 3 hours
        "medium": (350, 700),     # 3-7 hours  
        "long": (600, 1500),      # > 7 hours
    }
    
    # Estimate distance category based on common routes
    long_haul_dests = ["dubai", "tokyo", "sydney", "new york", "los angeles", "singapore", "bangkok", "bali"]
    medium_dests = ["paris", "rome", "barcelona", "amsterdam", "berlin", "madrid", "lisbon", "athens"]
    
    dest_lower = destination.lower()
    if any(d in dest_lower for d in long_haul_dests):
        min_p, max_p = base_prices["long"]
    elif any(d in dest_lower for d in medium_dests):
        min_p, max_p = base_prices["medium"]
    else:
        min_p, max_p = base_prices["short"]
    
    base = random.randint(min_p, max_p)
    
    # Cabin class multipliers
    if cabin_class == "business":
        base = int(base * 3.5)
    elif cabin_class == "first":
        base = int(base * 6)
    
    return base * passengers

@mcp.tool()
def search_flights(origin: str, destination: str, date: str, passengers: int = 1, return_date: str = "") -> str:
    """Search for flights between two cities. Returns a list of available flight options with booking links."""
    
    try:
        flight_date = datetime.strptime(date, "%Y-%m-%d")
        formatted_date = flight_date.strftime("%B %d, %Y")
    except ValueError:
        return "Error: Date must be in YYYY-MM-DD format."
    
    # Calculate return date if not provided (default 7 days)
    if not return_date:
        from datetime import timedelta
        return_dt = flight_date + timedelta(days=7)
        return_date = return_dt.strftime("%Y-%m-%d")
    
    # URL encode for booking links - clean city names
    origin_clean = origin.split(",")[0].strip()
    dest_clean = destination.split(",")[0].strip()
    
    # Different encoding formats for different sites
    origin_encoded = urllib.parse.quote(origin_clean)  # "San%20Francisco"
    dest_encoded = urllib.parse.quote(dest_clean)      # "Tokyo"
    origin_lower = origin_clean.lower().replace(" ", "-")  # "san-francisco"
    dest_lower = dest_clean.lower().replace(" ", "-")      # "tokyo"

    results = []
    results.append(f"✈️ **Flight Search: {origin} → {destination}**")
    results.append(f"📅 {formatted_date} | 👥 {passengers} passenger{'s' if passengers > 1 else ''}")
    results.append("")
    results.append("---")
    
    # Generate 4 realistic flight options
    used_airlines = random.sample(REAL_AIRLINES, min(4, len(REAL_AIRLINES)))
    
    flight_times = [
        (6, 30), (9, 15), (14, 45), (19, 20), (22, 10)
    ]
    random.shuffle(flight_times)
    
    for i, airline in enumerate(used_airlines[:4]):
        dep_hour, dep_min = flight_times[i]
        
        # Determine if direct or with layover
        is_direct = random.random() > 0.6  # 40% direct flights
        
        if is_direct:
            duration_h = random.randint(6, 8)
            duration_m = random.choice([0, 15, 30, 45])
            total_mins = duration_h * 60 + duration_m
            arr_hour = (dep_hour + duration_h + (dep_min + duration_m) // 60) % 24
            arr_min = (dep_min + duration_m) % 60
            next_day = "⁺¹" if dep_hour + duration_h >= 24 else ""
            stops_text = "Direct"
            layover_text = ""
        else:
            # Flight with layover
            layover_code = random.choice(list(LAYOVER_CITIES.keys()))
            layover_city = LAYOVER_CITIES[layover_code]
            layover_duration = random.randint(1, 3)
            layover_mins = random.choice([0, 15, 30, 45])
            
            leg1_duration = random.randint(3, 5)
            leg2_duration = random.randint(3, 5)
            total_duration_h = leg1_duration + layover_duration + leg2_duration
            total_mins = total_duration_h * 60 + layover_mins
            
            arr_hour = (dep_hour + total_duration_h) % 24
            arr_min = (dep_min + layover_mins) % 60
            next_day = "⁺¹" if dep_hour + total_duration_h >= 24 else ""
            stops_text = "1 stop"
            layover_text = f" via {layover_city} ({layover_duration}h {layover_mins}m layover)"
        
        # Price
        price = generate_realistic_price(origin, destination, passengers)
        flight_num = f"{airline['code']}{random.randint(100, 999)}"
        
        # CO2 estimate
        co2 = random.randint(400, 900)
        
        # Build booking URL for this specific flight - DON'T include airline name
        # Use simple city search for best results
        google_url = f"https://www.google.com/travel/flights?q={origin_encoded}%20to%20{dest_encoded}%20{date}%20to%20{return_date}"
        skyscanner_url = f"https://www.skyscanner.com/transport/flights/{origin_lower}/{dest_lower}/{date}/{return_date}/?adults={passengers}&cabinclass=economy"
        
        results.append("")
        results.append(f"### ✈️ Option {i+1}: {airline['name']}")
        results.append(f"**{flight_num}** • {stops_text}{layover_text}")
        results.append(f"🕐 {dep_hour:02d}:{dep_min:02d} → {arr_hour:02d}:{arr_min:02d}{next_day} ({total_mins // 60}h {total_mins % 60}m)")
        results.append(f"💰 **${price:,}** total for {passengers} passenger{'s' if passengers > 1 else ''}")
        results.append(f"🌱 {co2} kg CO₂")
        results.append(f"🔗 [Book on Google Flights]({google_url}) | [Book on Skyscanner]({skyscanner_url})")
        results.append("")
    
    results.append("---")
    results.append("")
    results.append("💡 **Tip:** Prices shown are estimates. Click booking links for live prices and availability.")
    
    return "\n".join(results)

@mcp.tool()
def get_flight_details(flight_number: str) -> str:
    """Get detailed information about a specific flight."""
    airline_code = flight_number[:2].upper()
    airline_name = next((a["name"] for a in REAL_AIRLINES if a["code"] == airline_code), "Unknown Airline")
    
    return f"""✈️ **Flight {flight_number} Details**

🛩️ **Airline:** {airline_name}
✈️ **Aircraft:** Boeing 787-9 Dreamliner
🚪 **Terminal:** {random.randint(1, 5)} | Gate: {random.choice(['A', 'B', 'C', 'D'])}{random.randint(1, 30)}
✅ **Status:** On Time

**Included Amenities:**
• 🧳 Checked baggage: 23kg
• 🎒 Carry-on: 7kg  
• 🍽️ Meal service included
• 📺 In-flight entertainment
• 🔌 USB charging ports

**Check-in:** Opens 24 hours before departure"""

if __name__ == "__main__":
    mcp.run()