ClearML downloads/caches datasets to ~/.clearml/ folder so yes, you need to modify your code.
dataset_folder = Dataset.get(project_name=, dataset_name=, version=).get_local_copy() file_json_path = os.path.join(dataset_folder, 'file.json')