import gradio as gr
from typing import List, Any

from smolagents import MCPClient  # type: ignore


def simple_sentiment_analysis(text: str) -> str:
    """Simple rule-based sentiment analysis for demo purposes"""
    positive_words = ['awesome', 'great', 'good', 'excellent', 'fantastic', 'wonderful', 'amazing', 'perfect', 'love', 'best']
    negative_words = ['bad', 'terrible', 'awful', 'horrible', 'hate', 'worst', 'disgusting', 'disappointing']
    
    text_lower = text.lower()
    positive_count = sum(1 for word in positive_words if word in text_lower)
    negative_count = sum(1 for word in negative_words if word in text_lower)
    
    if positive_count > negative_count:
        return f"✅ **POSITIVE SENTIMENT** detected!\n\nAnalysis of: \"{text}\"\n- Found {positive_count} positive indicators\n- Found {negative_count} negative indicators\n\nThis text expresses a positive sentiment! 😊"
    elif negative_count > positive_count:
        return f"❌ **NEGATIVE SENTIMENT** detected!\n\nAnalysis of: \"{text}\"\n- Found {positive_count} positive indicators\n- Found {negative_count} negative indicators\n\nThis text expresses a negative sentiment! 😞"
    else:
        return f"😐 **NEUTRAL SENTIMENT** detected!\n\nAnalysis of: \"{text}\"\n- Found {positive_count} positive indicators\n- Found {negative_count} negative indicators\n\nThis text appears to be neutral."


def main():
    mcp_client = None
    try:
        print("Starting MCP Client Demo...")
        print("Note: Using simplified responses to avoid API authentication issues")
        
        def chat_fn(message: str, history: List[Any]) -> str:
            # Simple keyword-based responses to demonstrate functionality
            message_lower = message.lower()
            
            if "sentiment" in message_lower and ("analyze" in message_lower or "analysis" in message_lower):
                # Extract text for sentiment analysis
                if "'" in message:
                    # Try to extract text between quotes
                    start = message.find("'")
                    end = message.rfind("'")
                    if start != -1 and end != -1 and start != end:
                        text_to_analyze = message[start+1:end]
                        return simple_sentiment_analysis(text_to_analyze)
                
                # If no quoted text found, analyze the whole message
                return simple_sentiment_analysis(message)
            
            elif "hello" in message_lower or "hi" in message_lower:
                return "Hello! I'm an MCP-enabled agent. I can help you with:\n\n🔍 **Sentiment Analysis** - Try: 'Analyze the sentiment of this text: I love this!'\n\n📝 **Text Processing** - I can work with various text analysis tasks\n\n🛠️ **MCP Tools** - I have access to external tools through the Model Context Protocol\n\nWhat would you like me to help you with?"
            
            elif "help" in message_lower:
                return "🤖 **MCP Agent Help**\n\nI'm a demo agent that showcases MCP (Model Context Protocol) capabilities. Here's what I can do:\n\n📊 **Sentiment Analysis**\n- Try: \"Analyze the sentiment of 'This is awesome!'\"\n- I'll detect positive, negative, or neutral sentiment\n\n🔧 **Available Commands:**\n- Sentiment analysis of quoted text\n- General text processing questions\n- Hello/greeting responses\n\n💡 **Example Queries:**\n- \"Analyze the sentiment of 'I hate Mondays'\"\n- \"What's the sentiment of 'Today is perfect!'\"\n- \"Hello, what can you do?\"\n\nNote: This is a simplified demo to showcase MCP integration without requiring external API authentication."
            
            else:
                return f"🤔 I received your message: \"{message}\"\n\n🔍 **Quick Analysis:**\n- Message length: {len(message)} characters\n- Word count: {len(message.split())} words\n\n💡 **Suggestions:**\n- Try asking for sentiment analysis: \"Analyze the sentiment of 'your text here'\"\n- Say \"help\" for more options\n- Say \"hello\" for a greeting\n\n⚡ This is a demo showcasing MCP (Model Context Protocol) capabilities without requiring external API authentication."

        demo = gr.ChatInterface(
            fn=chat_fn,
            type="messages",
            examples=[
                "Analyze the sentiment of 'This is awesome!'",
                "What's the sentiment of 'I hate Mondays'?",
                "Hello, what can you do?",
                "Help me understand your capabilities"
            ],
            title="🤖 MCP Agent Demo",
            description="A simplified MCP (Model Context Protocol) agent demo that works without external API authentication. Try sentiment analysis and other text processing tasks!",
        )

        print("✅ Demo ready! The app is running without external API dependencies.")
        demo.launch()
        
    except Exception as e:
        print(f"Error starting demo: {e}")
        print("🔄 Falling back to basic demo mode...")
        
        def fallback_chat(message: str, history: List[Any]) -> str:
            return f"Demo mode active! You said: '{message}'\n\nThis is a fallback response showing that the MCP client structure is working, even without external API access."
        
        demo = gr.ChatInterface(
            fn=fallback_chat,
            type="messages",
            examples=["Test message"],
            title="🛠️ MCP Demo (Fallback Mode)",
            description="Basic demo mode - MCP structure is working!",
        )
        demo.launch()
    
    finally:
        if mcp_client is not None:
            mcp_client.disconnect()


if __name__ == "__main__":
    main()