from transformers import pipeline
from PIL import Image
import numpy as np
import cv2  # OpenCV for better mask processing

# Initialize segmentation pipeline
segmenter = pipeline(model="mattmdjaga/segformer_b2_clothes")

def segment_clothing(img, clothes=["Hat", "Upper-clothes", "Skirt", "Pants", "Dress", "Belt", "Left-shoe", "Right-shoe", "Scarf"]):
    # Segment image
    segments = segmenter(img)

    # Create list of masks
    mask_list = []
    for s in segments:
        if s['label'] in clothes:
            mask_list.append(np.array(s['mask'], dtype=np.uint8))  # Convert to numpy array and ensure it's uint8

    # Initialize final mask with zeros
    final_mask = np.zeros_like(mask_list[0], dtype=np.uint8)

    # Combine masks into one
    for mask in mask_list:
        final_mask = np.maximum(final_mask, mask)

    # Optional: Dilate the mask to ensure coverage at edges
    kernel = np.ones((5, 5), np.uint8)
    final_mask = cv2.dilate(final_mask, kernel, iterations=2)

    # Optional: Erode to slightly smoothen the mask
    final_mask = cv2.erode(final_mask, kernel, iterations=1)

    # Optional: Use contour filling to ensure all areas within contours are filled
    contours, _ = cv2.findContours(final_mask, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)
    cv2.drawContours(final_mask, contours, -1, (255), thickness=cv2.FILLED)

    # Apply Gaussian blur to smooth edges and reduce noise
    final_mask = cv2.GaussianBlur(final_mask, (7, 7), 0)

    # Convert mask to binary (0 or 255) if needed for alpha channel
    _, final_mask = cv2.threshold(final_mask, 127, 255, cv2.THRESH_BINARY)

    # Convert final mask from np array to PIL image
    final_mask = Image.fromarray(final_mask)

    # Apply mask to original image (convert to RGBA first)
    img = img.convert("RGBA")
    img.putalpha(final_mask)

    return img