app.py

import os
from groq import Groq
import gradio as gr
from PIL import Image
from transformers import BlipProcessor, BlipForConditionalGeneration

# Retrieve Groq API Key securely from environment variables
api_key = os.getenv("GROQ_API_KEY")
if not api_key:
    raise ValueError("GROQ_API_KEY not found. Please add it to your environment variables.")

# Initialize Groq Client
client = Groq(api_key=api_key)

# Initialize BLIP model for image captioning
processor = BlipProcessor.from_pretrained("Salesforce/blip-image-captioning-base")
model = BlipForConditionalGeneration.from_pretrained("Salesforce/blip-image-captioning-base")

# Function to process image and generate caption
def process_image(image):
    inputs = processor(image, return_tensors="pt")
    outputs = model.generate(**inputs)
    caption = processor.decode(outputs[0], skip_special_tokens=True)
    return caption

# Function to generate chatbot responses with text and image
def chatbot_response(user_input, image=None, chat_history=[]):
    try:
        # Process image if provided
        image_caption = ""
        if image:
            image_caption = process_image(image)
        
        # Prepare input messages for Groq
        combined_input = user_input
        if image_caption:
            combined_input += f"\n\nImage Caption: {image_caption}"
        
        messages = [{"role": "user", "content": combined_input}]
        
        # Generate response from Groq API
        chat_completion = client.chat.completions.create(
            messages=messages,
            model="llama-3.3-70b-versatile",
            stream=False,
        )
        response = chat_completion.choices[0].message.content

        # Update chat history
        chat_history.append({"role": "user", "content": user_input})
        if image_caption:
            chat_history.append({"role": "user", "content": f"Image Caption: {image_caption}"})
        chat_history.append({"role": "assistant", "content": response})
        return chat_history, chat_history
    except Exception as e:
        return chat_history + [{"role": "assistant", "content": f"Error: {str(e)}"}], chat_history

# Create Gradio Interface with enhanced UI
with gr.Blocks(title="Chatbot with Text and Image Input") as interface:
    gr.Markdown(
        """
        # 🤖 **Chatbot with Text and Image Input**
        Powered by **Groq** and enhanced with **BLIP Image Captioning**. 
        
        📝 *Type your message below, attach a photo (optional), and get a response!*
        """
    )
    
    with gr.Row():
        with gr.Column(scale=4):
            chatbot = gr.Chatbot(label="Chat", type="messages")
        with gr.Column(scale=1):
            gr.Markdown(
                """
                ## Instructions
                1. Enter your message in the text box.
                2. Attach an image if needed.
                3. Click 'Send' or press 'Enter' to chat.
                4. Enjoy your enhanced conversation! 🎉
                """
            )
    
    user_input = gr.Textbox(
        label="Your Message",
        placeholder="Type your message here...",
        lines=1,
    )
    user_image = gr.Image(
        label="Attach Image (Optional)",
        type="pil",
    )
    submit_button = gr.Button("Send 🚀")
    clear_button = gr.Button("Clear Chat 🧹")
    
    chat_history = gr.State([])  # Initialize chat history
    
    # Define interaction logic
    submit_button.click(
        chatbot_response, 
        inputs=[user_input, user_image, chat_history], 
        outputs=[chatbot, chat_history],
    )
    clear_button.click(
        lambda: ([], []), 
        inputs=None, 
        outputs=[chatbot, chat_history],
    )

# Launch Gradio App
if __name__ == "__main__":
    interface.launch()