""" Author: Mélanie Gaillochet Date: 2020-10-27 """ import json import os import pickle as pkl import h5py import nibabel as nib import numpy as np from Configs.configs import config_folder def get_config_from_json(config_filename): """ Get the config from a json file :param config_filename: name of config file (located in Configs.config) :return: config(namespace) or config(dictionary) """ config_filepath = os.path.join(config_folder, config_filename) config_dict = _read_json_file(config_filepath) return config_dict def _read_json_file(file_path): """ We are reading the json file and returning a dictionary :param file_path: :return: """ # We parse the configurations from the config json file provided with open(file_path, 'r') as config_file: output_dict = json.load(config_file) return output_dict def load_single_image(folder_path, filename_list, idx): """ We load the data and label for the specific list index given :param folder_path: :param filename_list: :param idx: :return: image (array) """ cur_volume_path = os.path.join(folder_path, filename_list[idx]) ending = cur_volume_path.rpartition('.')[2] if ending == 'nii': inputImage = nib.load(cur_volume_path) img = inputImage.get_data() img = np.array(img) elif not os.path.isdir(folder_path): with h5py.File(folder_path + '.hdf5', 'r') as hf: img = hf[filename_list[idx]][:] return img def save_hdf5(data, img_idx, dest_file): """ We are saving an hdf5 object :param data: :param filename: :return: """ with h5py.File(dest_file, "a", libver='latest', swmr=True) as hf: hf.swmr_mode = True hf.create_dataset(name=str(img_idx), data=data, shape=data.shape, dtype=data.dtype) def create_unexisting_folder(dir_path): """ We create a folder with the given path. If the folder already exists, we add '_1', '_2', ... to it :param dir_path: """ i = 0 created = False path = dir_path while not created: try: os.makedirs(path) created = True except OSError or FileExistsError: i += 1 path = dir_path + '_' + str(i) # print(path) return path def save_obj(obj, name): """ Shortcut function to save an object as pkl Args: obj: object to save name: filename of the object """ with open(name + '.pkl', 'wb') as f: pkl.dump(obj, f, pkl.HIGHEST_PROTOCOL) def load_obj(name): """ Shortcut function to load an object from pkl file Args: name: filename of the object Returns: obj: object to load """ with open(name + '.pkl', 'rb') as f: return pkl.load(f) def unpickle(file): with open(file, 'rb') as fo: dict1 = pkl.load(fo, encoding='bytes') return dict1