app.py

import gradio as gr
from PIL import Image
import os
from utils import describe_image, text_to_speech, test_tts

print("🚀 Initializing Blind Vision Assistant...")

# Test TTS on startup
print("🧪 Testing audio system...")
tts_working = test_tts()
if tts_working:
    print("✅ Audio system is working!")
else:
    print("⚠️ Audio system may have issues")

class BlindVisionAssistant:
    def __init__(self):
        self.is_processing = False
        
    def process_surroundings(self, image):
        """
        Process the image and return audio description
        """
        print("📸 Processing image...")
        
        if image is None:
            print("❌ No image provided")
            return None, "Please upload an image first.", "⚠️ No image provided"
            
        try:
            self.is_processing = True
            print("🔄 Generating description...")
            
            # Get description from vision model
            description = describe_image(image)
            print(f"📝 Description: {description}")
            
            # Convert description to speech
            print("🔊 Converting to speech...")
            audio_path = text_to_speech(description)
            
            if audio_path and os.path.exists(audio_path):
                file_size = os.path.getsize(audio_path)
                print(f"✅ Audio file created: {audio_path} ({file_size} bytes)")
                self.is_processing = False
                return audio_path, f"🎯 {description}", "✅ Description ready! Click play to listen."
            else:
                print("❌ Audio file was not created")
                self.is_processing = False
                return None, f"🎯 {description}", "⚠️ Description generated but audio failed. Read the text above."
            
        except Exception as e:
            self.is_processing = False
            print(f"❌ Error: {str(e)}")
            error_msg = f"❌ Error: {str(e)}"
            return None, error_msg, "❌ Processing failed"

def create_interface():
    assistant = BlindVisionAssistant()
    
    with gr.Blocks(title="Blind Vision Assistant - See with Sound") as demo:
        
        # Header Section with Credits
        gr.Markdown("""
        # 👁️‍🗨️ Blind Vision Assistant
        ### *Helping visually impaired people understand their surroundings through sound*
        
        **Created by: Pritam Mahesh Phalake**  
        *Under the guidance of: Seema Uttam Jangam*
        
        ---
        """)
        
        # Instructions
        with gr.Row():
            with gr.Column():
                gr.Markdown("""
                ### 📋 How to Use:
                1. **Upload an image** using the upload button below
                2. Click the **'Describe Image'** button
                3. **Listen** to the audio description
                4. **Read** the text description below
                
                *Note: Audio may take a few seconds to generate*
                """)
        
        # Main Content
        with gr.Row():
            # Left Column - Input
            with gr.Column():
                gr.Markdown("### 📷 Image Input")
                
                image_input = gr.Image(
                    label="Upload Image",
                    type="filepath",
                    interactive=True,
                    height=300
                )
                
                with gr.Row():
                    process_btn = gr.Button(
                        "🎤 Describe Image",
                        variant="primary",
                        size="lg"
                    )
                    
                    clear_btn = gr.Button(
                        "🔄 Clear",
                        variant="secondary"
                    )
            
            # Right Column - Output
            with gr.Column():
                gr.Markdown("### 🔊 Audio Output")
                
                audio_output = gr.Audio(
                    label="Audio Description",
                    interactive=False,
                    type="filepath"
                )
                
                gr.Markdown("### 📝 Text Description")
                
                text_output = gr.Textbox(
                    label="Image Description",
                    placeholder="Description will appear here after processing...",
                    lines=4,
                    interactive=False
                )
                
                status_output = gr.Textbox(
                    label="Status",
                    value="🟢 Ready - Upload an image above",
                    interactive=False
                )
        
        gr.Markdown("---")
        gr.Markdown("""
        <div style='text-align: center'>
        <p>Built with ❤️ for accessibility | Uses BLIP AI</p>
        </div>
        """)
        
        # Event Handlers
        def clear_all():
            return None, None, None, "🟢 Ready - Upload an image above"
        
        # Connect buttons
        process_btn.click(
            fn=assistant.process_surroundings,
            inputs=[image_input],
            outputs=[audio_output, text_output, status_output]
        )
        
        clear_btn.click(
            fn=clear_all,
            inputs=[],
            outputs=[image_input, audio_output, text_output, status_output]
        )
        
    return demo

if __name__ == "__main__":
    print("🚀 Starting Blind Vision Assistant...")
    print("📸 Please upload an image to get started")
    
    demo = create_interface()
    demo.launch(
        server_name="0.0.0.0",
        server_port=7860
    )