data/test_main.py

#!/usr/bin/env python
"""
For evaluation
Extended from ADNet code by Hansen et al.
"""
import shutil
import SimpleITK as sitk
import torch.backends.cudnn as cudnn
import torch.optim
from torch.utils.data import DataLoader
from models.fewshot import FewShotSeg
from dataloaders.datasets import TestDataset
from dataloaders.dataset_specifics import *
from utils import *
from config import ex


@ex.automain
def main(_run, _config, _log):
    if _run.observers:
        os.makedirs(f'{_run.observers[0].dir}/interm_preds', exist_ok=True)
        for source_file, _ in _run.experiment_info['sources']:
            os.makedirs(os.path.dirname(f'{_run.observers[0].dir}/source/{source_file}'),
                        exist_ok=True)
            _run.observers[0].save_file(source_file, f'source/{source_file}')
        shutil.rmtree(f'{_run.observers[0].basedir}/_sources')

        # Set up logger -> log to .txt
        file_handler = logging.FileHandler(os.path.join(f'{_run.observers[0].dir}', f'logger.log'))
        file_handler.setLevel('INFO')
        formatter = logging.Formatter('%(asctime)s - %(levelname)s - %(name)s - %(message)s')
        file_handler.setFormatter(formatter)
        _log.handlers.append(file_handler)
        _log.info(f'Run "{_config["exp_str"]}" with ID "{_run.observers[0].dir[-1]}"')

    # Deterministic setting for reproduciablity.
    if _config['seed'] is not None:
        random.seed(_config['seed'])
        torch.manual_seed(_config['seed'])
        torch.cuda.manual_seed_all(_config['seed'])
        cudnn.deterministic = True

    # Enable cuDNN benchmark mode to select the fastest convolution algorithm.
    cudnn.enabled = True
    cudnn.benchmark = True
    torch.cuda.set_device(device=_config['gpu_id'])
    torch.set_num_threads(1)

    _log.info(f'Create model...')
    model = FewShotSeg(alpha=_config['alpha'])
    model.cuda()
    model.load_state_dict(torch.load(_config['reload_model_path'], map_location='cpu'))

    _log.info(f'Load data...')
    data_config = {
        'data_dir': _config['path'][_config['dataset']]['data_dir'],
        'dataset': _config['dataset'],
        'n_shot': _config['n_shot'],
        'n_way': _config['n_way'],
        'n_query': _config['n_query'],
        'n_sv': _config['n_sv'],
        'max_iter': _config['max_iters_per_load'],
        'eval_fold': _config['eval_fold'],
        'min_size': _config['min_size'],
        'max_slices': _config['max_slices'],
        'supp_idx': _config['supp_idx'],
    }
    test_dataset = TestDataset(data_config)
    test_loader = DataLoader(test_dataset,
                             batch_size=_config['batch_size'],
                             shuffle=False,
                             num_workers=_config['num_workers'],
                             pin_memory=True,
                             drop_last=True)

    # Get unique labels (classes).
    labels = get_label_names(_config['dataset'])

    # Loop over classes.
    class_dice = {}
    class_iou = {}

    _log.info(f'Starting validation...')
    for label_val, label_name in labels.items():

        # Skip BG class.
        if label_name == 'BG':
            continue
        elif (not np.intersect1d([label_val], _config['test_label'])):
            continue

        _log.info(f'Test Class: {label_name}')

        # Get support sample + mask for current class.
        support_sample = test_dataset.getSupport(label=label_val, all_slices=False, N=_config['n_part'])
        test_dataset.label = label_val

        # Test.
        with torch.no_grad():
            model.eval()

            # Unpack support data.
            support_image = [support_sample['image'][[i]].float().cuda() for i in
                             range(support_sample['image'].shape[0])]  # n_shot x 3 x H x W
            support_fg_mask = [support_sample['label'][[i]].float().cuda() for i in
                               range(support_sample['image'].shape[0])]  # n_shot x H x W

            # Loop through query volumes.
            scores = Scores()
            for i, sample in enumerate(test_loader):

                # Unpack query data.
                query_image = [sample['image'][i].float().cuda() for i in
                               range(sample['image'].shape[0])]  # [C x 3 x H x W]
                query_label = sample['label'].long()  # C x H x W
                query_id = sample['id'][0].split('image_')[1][:-len('.nii.gz')]

                # Compute output.
                # Match support slice and query sub-chunck.
                query_pred = torch.zeros(query_label.shape[-3:])
                C_q = sample['image'].shape[1]
                idx_ = np.linspace(0, C_q, _config['n_part'] + 1).astype('int')
                for sub_chunck in range(_config['n_part']):
                    support_image_s = [support_image[sub_chunck]]  # 1 x 3 x H x W
                    support_fg_mask_s = [support_fg_mask[sub_chunck]]  # 1 x H x W
                    query_image_s = query_image[0][idx_[sub_chunck]:idx_[sub_chunck + 1]]  # C' x 3 x H x W
                    query_pred_s = []
                    for i in range(query_image_s.shape[0]):
                        _pred_s, _ = model([support_image_s], [support_fg_mask_s], [query_image_s[[i]]],
                                           train=False, n_iters=_config['n_iters'])  # C x 2 x H x W
                        query_pred_s.append(_pred_s)
                    query_pred_s = torch.cat(query_pred_s, dim=0)
                    query_pred_s = query_pred_s.argmax(dim=1).cpu()  # C x H x W
                    query_pred[idx_[sub_chunck]:idx_[sub_chunck + 1]] = query_pred_s

                # Record scores.
                scores.record(query_pred, query_label)

                # Log.
                _log.info(
                    f'Tested query volume: {sample["id"][0][len(_config["path"][_config["dataset"]]["data_dir"]):]}.')
                _log.info(f'Dice score: {scores.patient_dice[-1].item()}')

                # Save predictions.
                file_name = os.path.join(f'{_run.observers[0].dir}/interm_preds',
                                         f'prediction_{query_id}_{label_name}.nii.gz')
                itk_pred = sitk.GetImageFromArray(query_pred)
                sitk.WriteImage(itk_pred, file_name, True)
                _log.info(f'{query_id} has been saved. ')

            # Log class-wise results
            class_dice[label_name] = torch.tensor(scores.patient_dice).mean().item()
            class_iou[label_name] = torch.tensor(scores.patient_iou).mean().item()
            _log.info(f'Test Class: {label_name}')
            _log.info(f'Mean class IoU: {class_iou[label_name]}')
            _log.info(f'Mean class Dice: {class_dice[label_name]}')

    _log.info(f'Final results...')
    _log.info(f'Mean IoU: {class_iou}')
    _log.info(f'Mean Dice: {class_dice}')

    _log.info(f'End of validation.')
    return 1