app.py

"""
INVINCIX Chatbot - Gradio Interface with Hugging Face Model
Beautiful UI with chat history and settings
"""

import gradio as gr
from transformers import AutoModelForCausalLM, AutoTokenizer, BitsAndBytesConfig
import torch

# ================================================================
# CONFIGURATION
# ================================================================

# TODO: Replace with your Hugging Face model path
MODEL_NAME = "amanjain96/invincix-chatbot"  # Your HF model
TITLE = "💬 INVINCIX Chatbot"
DESCRIPTION = """
**"Simplicity is our culture and simplification is what we do"**

Welcome to INVINCIX! I'm here to help you learn about our services, products, and how we can support your business transformation journey.

Ask me about:
- 🛠️ Our Services (Software Development, Product Engineering, Mobile Apps)
- 📦 Our Products (Xprodedge, DigiStack, Zikshaa, AD4P)
- 🌟 Company Culture & Values
- 📍 Locations & Contact Information
"""

COMPANY_INFO = """
### 🎯 About INVINCIX

**Mission**: Digitize 500+ startups

**Core Values**:
- 🤝 Engage
- 💡 Innovate  
- 🚀 Invent
- ⭐ Excel

**Services**:
- Software Development
- Product Engineering
- Mobile Applications
- Data Analytics
- Cloud Solutions
- IoT & Drone Solutions

**Products**:
- **Xprodedge**: Product innovation platform
- **DigiStack**: Service automation
- **Zikshaa**: Mentorship platform
- **AD4P**: Application development

**Contact**:
- 📧 info@invincix.com
- 📱 +91 674 297 2316
- 🌐 [invincix.com](https://invincix.com)
- 📍 Bhubaneswar, India + 15 global locations
"""

# ================================================================
# LOAD MODEL
# ================================================================

device = "cuda" if torch.cuda.is_available() else "cpu"
print(f"Loading model on {device}...")




try:
    
    tokenizer = AutoTokenizer.from_pretrained(MODEL_NAME)
    model = AutoModelForCausalLM.from_pretrained(
        MODEL_NAME,
        torch_dtype=torch.float16 if device == "cuda" else torch.float32,
        device_map="auto" if device == "cuda" else None,
        low_cpu_mem_usage=True
    )
    
    if device == "cpu":
        model = model.to(device)
    
    print("✅ Model loaded successfully!")
    MODEL_LOADED = True
except Exception as e:
    print(f"❌ Error loading model: {e}")
    MODEL_LOADED = False
    error_message = str(e)

# ================================================================
# GENERATE RESPONSE
# ================================================================

def generate_response(message, history, temperature=0.7, max_tokens=300):
    """Generate chatbot response"""
    
    if not MODEL_LOADED:
        return f"❌ Model failed to load. Error: {error_message}\n\nPlease check your model path and try again."
    
    # Build prompt with history
    conversation = ""
    for user_msg, bot_msg in history:
        conversation += f"User: {user_msg}\nAssistant: {bot_msg}\n\n"
    
    prompt = f"""Below is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request as an INVINCIX chatbot assistant.

### Instruction:
You are a helpful and empathetic assistant for INVINCIX, a software engineering and product development company. Respond to the customer's inquiry with warmth, professionalism, and accurate information about INVINCIX's services, products, and values. Keep your response concise and engaging.

### Previous Conversation:
{conversation}

### Input:
{message}

### Response:
"""
    
    try:
        # Tokenize
        inputs = tokenizer(prompt, return_tensors="pt").to(device)
        
        # Generate
        with torch.no_grad():
            outputs = model.generate(
                **inputs,
                max_new_tokens=max_tokens,
                temperature=temperature,
                top_p=0.9,
                do_sample=True,
                pad_token_id=tokenizer.eos_token_id,
                eos_token_id=tokenizer.eos_token_id,
            )
        
        # Decode
        full_response = tokenizer.decode(outputs[0], skip_special_tokens=True)
        
        # Extract response
        if "### Response:" in full_response:
            answer = full_response.split("### Response:")[-1].strip()
        else:
            answer = full_response[len(prompt):].strip()
        
        # Clean up
        answer = answer.split("###")[0].strip()
        answer = answer.split("### Input:")[0].strip()
        
        return answer
    
    except Exception as e:
        return f"I apologize, but I encountered an error: {str(e)}. Please try again."

# ================================================================
# GRADIO INTERFACE
# ================================================================

# Custom CSS for INVINCIX branding
custom_css = """
#title {
    background: linear-gradient(135deg, #667eea 0%, #764ba2 100%);
    padding: 20px;
    border-radius: 10px;
    color: white;
    text-align: center;
    margin-bottom: 20px;
}

#chatbot {
    height: 500px;
}

.gradio-container {
    font-family: 'Arial', sans-serif;
}

#footer {
    text-align: center;
    color: #7f8c8d;
    padding: 20px;
    font-size: 0.9em;
}
"""

# Create interface
with gr.Blocks(css=custom_css, theme=gr.themes.Soft()) as demo:
    
    # Header
    gr.Markdown(
        f"""
        <div id="title">
            <h1>{TITLE}</h1>
        </div>
        """
    )
    
    gr.Markdown(DESCRIPTION)
    
    # Main chat interface
    with gr.Row():
        with gr.Column(scale=3):
            chatbot = gr.Chatbot(
                elem_id="chatbot",
                label="Chat with INVINCIX",
                avatar_images=(None, "🤖"),
                bubble_full_width=False,
                show_copy_button=True
            )
            
            with gr.Row():
                msg = gr.Textbox(
                    label="Your Message",
                    placeholder="Ask me anything about INVINCIX...",
                    lines=2,
                    scale=4
                )
                submit = gr.Button("Send 📤", variant="primary", scale=1)
            
            with gr.Row():
                clear = gr.Button("🗑️ Clear Chat")
                retry = gr.Button("🔄 Retry")
        
        # Sidebar with settings and info
        with gr.Column(scale=1):
            gr.Markdown("### ⚙️ Settings")
            
            temperature = gr.Slider(
                minimum=0.1,
                maximum=1.0,
                value=0.7,
                step=0.1,
                label="Temperature",
                info="Lower = more consistent, Higher = more creative"
            )
            
            max_tokens = gr.Slider(
                minimum=100,
                maximum=500,
                value=300,
                step=50,
                label="Max Tokens",
                info="Maximum response length"
            )
            
            gr.Markdown("---")
            gr.Markdown(COMPANY_INFO)
            
            # System status
            status_emoji = "⚡" if device == "cuda" else "💻"
            model_status = "✅ Loaded" if MODEL_LOADED else "❌ Error"
            
            gr.Markdown(f"""
            ### 📊 System Status
            - Device: {status_emoji} {device.upper()}
            - Model: {model_status}
            - Ready: {"✅" if MODEL_LOADED else "❌"}
            """)
    
    # Examples
    gr.Examples(
        examples=[
            "What is INVINCIX?",
            "Tell me about your services",
            "What is Xprodedge?",
            "How can you help startups?",
            "What are your office locations?",
            "Tell me about your company culture",
        ],
        inputs=msg,
        label="💡 Try these questions:"
    )
    
    # Footer
    gr.Markdown(
        """
        <div id="footer">
            <p><em>"We keep our feet grounded to ensure your head is in the cloud"</em></p>
            <p>Powered by fine-tuned AI | Built for INVINCIX</p>
        </div>
        """
    )
    
    # Event handlers
    def respond(message, chat_history, temp, max_tok):
        if not message.strip():
            return "", chat_history
        
        bot_message = generate_response(message, chat_history, temp, max_tok)
        chat_history.append((message, bot_message))
        return "", chat_history
    
    def retry_last(chat_history, temp, max_tok):
        if not chat_history:
            return chat_history
        
        last_message = chat_history[-1][0]
        chat_history = chat_history[:-1]
        bot_message = generate_response(last_message, chat_history, temp, max_tok)
        chat_history.append((last_message, bot_message))
        return chat_history
    
    # Wire up events
    msg.submit(respond, [msg, chatbot, temperature, max_tokens], [msg, chatbot])
    submit.click(respond, [msg, chatbot, temperature, max_tokens], [msg, chatbot])
    clear.click(lambda: None, None, chatbot, queue=False)
    retry.click(retry_last, [chatbot, temperature, max_tokens], chatbot)

# ================================================================
# LAUNCH
# ================================================================

if __name__ == "__main__":
    demo.launch(
        server_name="0.0.0.0",  # Allow external access
        server_port=7860,        # Default Gradio port
        share=False,             # Set to True to create public link
        show_error=True
    )