import os

import numpy as np
import pandas as pd
import SimpleITK as sitk

from skimage.measure import label, regionprops

from glob import glob
from tqdm import tqdm


masks = sorted(glob('luna25_ts_seg/*.nii.gz'))

record = []
for mask_file in tqdm(masks):

    mask_itk = sitk.ReadImage(mask_file)
    mask_arr = sitk.GetArrayFromImage(mask_itk)

    unit_volume = np.prod(mask_itk.GetSpacing())

    binary_mask = (mask_arr == 2).astype(int)  # Extract nodule label

    labeled_mask = label(binary_mask, connectivity=3)  # 3D connectivity
    properties = regionprops(labeled_mask)

    for prop in properties:
        # print(f"Region {prop.label}: Centroid = {prop.centroid}, Volume = {prop.area} voxels")

        record.append({
            'mask_file': os.path.basename(mask_file),
            'centroid': np.round(prop.centroid, 3),
            'volume_px': int(prop.area),
            'volume_mm3': np.round(prop.area * unit_volume, 3),
            'bbox': np.array(prop.bbox),
            # 'axis_major_length': prop.axis_major_length,
            # 'axis_minor_length': prop.axis_minor_length
            })

record_df = pd.DataFrame(record)
record_df.to_csv('metadata.csv', index=False)