"""
Project: Customer Service Bot
Author: RSK World
Website: https://rskworld.in
Contact: info@rskworld.com, support@rskworld.com
Phone: +91 93305 39277
Year: 2026
"""

from difflib import get_close_matches

# Extended Knowledge Base
FAQ_KB = {
    "hours": "We are open 24/7 for your convenience!",
    "location": "We are a digital-first company, operating globally.",
    "refund": "Refunds are processed within 3-5 business days. Please provide your Order ID.",
    "pricing": "Our basic plan starts at $0. Free tier available. Premium plans start at $10/mo.",
    "hiring": "We are always looking for talent! Check our careers page.",
    "contact": "You can reach us at info@rskworld.com or +91 93305 39277.",
    "hello": "Hello! Welcome to our Customer Service. How can I help you today?",
    "hi": "Hi there! Need assistance?",
    "bye": "Goodbye! Have a great day ahead.",
    "password": "To reset your password, click 'Forgot Password' on the login page.",
    "shipping": "Shipping usually takes 2-4 business days within the country.",
    "payment": "We accept Visa, Mastercard, PayPal, and Crypto.",
    "cancel": "You can cancel your subscription anytime from your settings.",
    "agent": "I can connect you to a human agent if you raise a ticket.",
    "status": "You can check the status of your ticket on the 'My Tickets' page.",
    "features": "Our bot supports FAQ handling, ticket management, and sentiment analysis."
}

# Negative words for basic sentiment analysis
NEGATIVE_WORDS = ['angry', 'bad', 'terrible', 'worst', 'fail', 'broken', 'scam', 'late', 'hate', 'stupid', 'useless', 'waiting', 'ridiculous']
URGENT_WORDS = ['manager', 'supervisor', 'emergency', 'urgent', 'legal', 'sue']

def analyze_sentiment(text):
    """
    Basic sentiment analysis to detect frustration and urgency.
    Returns: (score, priority_label)
    """
    text = text.lower()
    score = 0
    
    # Check for urgency keywords
    for word in URGENT_WORDS:
        if word in text:
            return 10, 'High' # Immediate High Priority

    for word in NEGATIVE_WORDS:
        if word in text:
            score += 1
            
    if score >= 2:
        return score, 'High'
    elif score == 1:
        return score, 'Medium'
    else:
        return score, 'Normal'

def get_response(message):
    """
    Determine the bot response using Fuzzy Matching, Intent Detection, and Sentiment Adaptation.
    Project: Customer Service Bot
    Author: RSK World
    Year: 2026
    """
    msg = message.lower().strip()
    sentiment_score, priority = analyze_sentiment(msg)
    
    # Adaptive Tone: Apologize if user is upset
    prefix = ""
    if sentiment_score > 0:
        prefix = "I understand you are upset. I apologize for the inconvenience. "

    # 1. Check for immediate escalation triggers
    if priority == 'High' and "ticket" not in msg:
        return {
            "text": f"{prefix}It sounds like this is urgent. I strongly recommend raising a high-priority ticket immediately.",
            "type": "option",
            "options": ["Yes, raise a ticket"]
        }

    # 2. Advanced Fuzzy Matching for FAQ
    keys = list(FAQ_KB.keys())
    msg_words = msg.split()
    best_match = None
    
    # Check words
    for word in msg_words:
        matches = get_close_matches(word, keys, n=1, cutoff=0.7)
        if matches:
            best_match = matches[0]
            break
            
    # Direct search
    if not best_match:
        for key in keys:
            if key in msg:
                best_match = key
                break
                
    if best_match:
        return {
            "text": f"{prefix}{FAQ_KB[best_match]}",
            "type": "text"
        }
    
    # 3. Intent Detection for Ticket/Support
    if "ticket" in msg or "issue" in msg or "problem" in msg or "help" in msg or "human" in msg:
        response_text = "It seems you have an issue."
        if priority == 'High':
            response_text = f"{prefix}Let's prioritize getting you help. Would you like to raise a high-priority ticket?"
        else:
            response_text = f"{prefix}Would you like to raise a support ticket so our team can help?"
            
        return {
            "text": response_text,
            "type": "option",
            "options": ["Yes, raise a ticket"]
        }
        
    # 4. Smart Fallback with Suggestion
    return {
        "text": f"{prefix}I'm not quite sure about that. I can help with FAQs (hours, refunds, pricing) or you can say 'help' to talk to support.",
        "type": "text"
    }