app.py

import gradio as gr
import cv2
import numpy as np
from model import TrafficSignDetector

# Load the detector
detector = TrafficSignDetector('config.yaml')

def detect_traffic_signs(image, confidence_threshold):
    """
    Process the uploaded image and return the image with detected signs.
    """
    # Validate input image
    if image is None:
        print("No image provided")
        return None, None
    
    print(f"Received image type: {type(image)}")
    if hasattr(image, 'convert'):
        image = np.array(image)
        print(f"Converted PIL to numpy array, shape: {image.shape}")
    
    # Check if image is valid
    if image.size == 0 or len(image.shape) != 3:
        print(f"Invalid image: shape={image.shape}")
        return None, None

    # Convert RGB to BGR for OpenCV
    image = cv2.cvtColor(image, cv2.COLOR_RGB2BGR)
    print(f"Converted to BGR, shape: {image.shape}")

    # Perform detection
    result_image, preprocessed_image = detector.detect(image, confidence_threshold=confidence_threshold)

    # Convert back to RGB for Gradio
    result_image = cv2.cvtColor(result_image, cv2.COLOR_BGR2RGB)
    preprocessed_image = cv2.cvtColor(preprocessed_image, cv2.COLOR_BGR2RGB)

    return result_image, preprocessed_image

# Create Gradio interface
with gr.Blocks(title="Traffic Sign Detector") as demo:
    gr.Markdown("# Traffic Sign Detector")
    gr.Markdown("Upload an image to detect traffic signs using YOLOv8. Detection runs automatically when you upload or adjust the threshold.")
    
    with gr.Row():
        input_image = gr.Image(label="Upload Image", type="pil")
        output_image = gr.Image(label="Detected Signs", interactive=False)
    
    with gr.Row():
        preprocessed_image = gr.Image(label="Preprocessed Image (640x640, Letterboxed)", interactive=False)
    
    with gr.Row():
        confidence_threshold = gr.Slider(
            minimum=0.01,
            maximum=0.9,
            value=0.30,
            step=0.01,
            label="Confidence Threshold",
            info="Lower values show more detections (less confident). Adjust to find optimal balance."
        )
    
    with gr.Row():
        detect_btn = gr.Button("Detect Traffic Signs", variant="primary")
        reset_btn = gr.Button("Clear")
    
    # Auto-detect on image upload
    input_image.change(
        fn=detect_traffic_signs,
        inputs=[input_image, confidence_threshold],
        outputs=[output_image, preprocessed_image],
        queue=True
    )
    
    # Auto-detect on threshold change
    confidence_threshold.change(
        fn=detect_traffic_signs,
        inputs=[input_image, confidence_threshold],
        outputs=[output_image, preprocessed_image],
        queue=True
    )
    
    # Manual detect button
    detect_btn.click(
        fn=detect_traffic_signs, 
        inputs=[input_image, confidence_threshold], 
        outputs=[output_image, preprocessed_image],
        queue=True
    )
    
    # Clear button
    reset_btn.click(
        fn=lambda: (None, None, None, 0.30),
        outputs=[input_image, output_image, preprocessed_image, confidence_threshold]
    )

if __name__ == "__main__":
    demo.queue().launch(server_name="0.0.0.0", server_port=7860)