docker_task_2/nnunetv2/nnsyn/nnsyn_preprocessing_seg.py

import glob
import os
import SimpleITK as sitk
import numpy as np
from concurrent.futures import ThreadPoolExecutor
from tqdm import tqdm
import json
import shutil

def copy_dataset_json(raw_path, preprocessed_path, dataset_name):
    src_path = os.path.join(raw_path, dataset_name, 'dataset.json')
    dst_path = os.path.join(preprocessed_path, dataset_name, 'dataset.json')
    if not os.path.exists(dst_path):
        shutil.copy(src_path, dst_path)
    else:
        print(f"Dataset JSON already exists at {dst_path}, skipping copy.")

def create_dataset_json(num_train, preprocessing, dataset_data_path, label_to_use): 
    labels = {str(label): i + 1 for i, label in enumerate(label_to_use)}
    labels["background"] = 0

    data_dataset_json = {
        "labels": labels,
        "channel_names": {
            "0": preprocessing,
            
        },
        "numTraining": num_train,
        "file_ending": ".mha"
    }
    dump_data_datasets_path = os.path.join(dataset_data_path, 'dataset.json')
    with open(dump_data_datasets_path, 'w') as f:
        json.dump(data_dataset_json, f)

def makedirs_raw_dataset(dataset_data_path):
    
    os.makedirs(dataset_data_path, exist_ok = True)
    os.makedirs(os.path.join(dataset_data_path, 'imagesTr'), exist_ok=True)
    os.makedirs(os.path.join(dataset_data_path, 'labelsTr'), exist_ok = True)

def process_file(data_path, dataset_path, modality_suffix="_0000"):
    curr_img = sitk.ReadImage(data_path)
    filename = os.path.basename(data_path)
    if not filename.endswith(f'{modality_suffix}.mha'):
        filename = filename + f'{modality_suffix}.mha'
    sitk.WriteImage(curr_img, os.path.join(dataset_path, f'imagesTr/{filename}'))

def process_file_labels(data_path, dataset_path):
    curr_img = sitk.ReadImage(data_path)
    filename = os.path.basename(data_path)
    sitk.WriteImage(curr_img, os.path.join(dataset_path, f'labelsTr/{filename}'))


def nnsyn_plan_and_preprocess_seg(data_origin_path: str, dataset_id: int,  dataset_id_src: int,
                        preprocessing_target: str,
                        configuration: str = '3d_fullres', plan: str = 'nnUNetPlans', 
                        dataset_name: str = None):
    list_data_ct = sorted(glob.glob(os.path.join(data_origin_path, 'TARGET_IMAGES','*.mha'), recursive=True))
    list_data_labels = sorted(glob.glob(os.path.join(data_origin_path, 'LABELS','*.mha'), recursive=True))
    print("target ---", len(list_data_ct), list_data_ct[:2])
    print("labels ---", len(list_data_labels), list_data_labels[:2])

    if dataset_name is None:
        dataset_name = 'SEG_' + os.path.basename(data_origin_path)


    # copy data from orign to nnUNet_raw
    dataset_data_path = os.path.join(os.environ['nnUNet_raw'], f'Dataset{dataset_id:03d}_{dataset_name}') 
    makedirs_raw_dataset(dataset_data_path)

    with ThreadPoolExecutor() as executor:
        list(tqdm(executor.map(lambda data_path: process_file(data_path, dataset_data_path, "_0000"), list_data_ct), total=len(list_data_ct)))

    with ThreadPoolExecutor() as executor:
        list(tqdm(executor.map(lambda target_path: process_file_labels(target_path, dataset_data_path), list_data_labels), total=len(list_data_labels)))

    
    # create dataset.json
    num_train = len(list_data_ct)
    assert len(list_data_labels) == len(list_data_ct)
    labels_to_use = [
                2, # kidney right
                3, # kidney left
                5, # liver
                6, # stomach
                *range(10, 14+1), #lungs
                *range(26, 50+1), #vertebrae
                51, #heart
                79, # spinal cord
                *range(92, 115+1), # ribs
                116 #sternum
            ]
    create_dataset_json(num_train, preprocessing_target, dataset_data_path, labels_to_use)
    
    SOURCE_PLAN_IDENTIFIER = plan
    TARGET_PLAN_IDENTIFIER = plan + f'_Dataset{dataset_id_src}'

    os.system(f'nnUNetv2_extract_fingerprint -d {dataset_id} --verify_dataset_integrity')
    os.system(f'nnUNetv2_move_plans_between_datasets -s {dataset_id_src} -t {dataset_id} -sp {SOURCE_PLAN_IDENTIFIER} -tp {TARGET_PLAN_IDENTIFIER}')
    copy_dataset_json(os.environ['nnUNet_raw'], os.environ['nnUNet_preprocessed'], f'Dataset{dataset_id:03d}_{dataset_name}')
    os.system(f'nnUNetv2_preprocess -d {dataset_id} -c {configuration} -plans_name {TARGET_PLAN_IDENTIFIER} -np 4')

if __name__ == '__main__':
    # example usage:
    # python -m nnsyn_preprocessing_entry -d 982 --data_origin_path '/datasets/work/hb-synthrad2023/work/synthrad2025/bw_workplace/data/nnunet_struct/ORIGIN/Synthrad2025_MRI2CT_AB' --preprocessing_target CT --dataset_id_src 960
    # nnsyn_plan_and_preprocess -d 982 --data_origin_path '/datasets/work/hb-synthrad2023/work/synthrad2025/bw_workplace/data/nnunet_struct/ORIGIN/Synthrad2025_MRI2CT_AB' --preprocessing_target CT --dataset_id_src 960
    nnsyn_plan_and_preprocess_seg(data_origin_path='/datasets/work/hb-synthrad2023/work/synthrad2025/bw_workplace/data/nnunet_struct/ORIGIN/synthrad2025_task1_mri2ct_AB', 
                              dataset_id=961, dataset_id_src=960,
                              preprocessing_target='CT', 
                              configuration='3d_fullres', plan='nnUNetResEncUNetLPlans')