run.py

"""Brain Mapper Run File

Description:

    This file contains all the relevant functions for running BrainMapper.
    The network can be ran in one of these modes:
        - train
        - evaluate path
        - evaluate whole

    TODO: Might be worth adding some information on uncertaintiy estimation, later down the line

Usage:

    In order to run the network, in the terminal, the user needs to pass it relevant arguments:

        $ ./setup.sh
        $ source env/bin/activate
        $ python run.py --mode ...

    The arguments for mode are the following:

        mode=train # For training the model
        mode=evaluate-score # For evaluating the model score
        mode=evaluate-mapping # For evaluating the model mapping
        # For clearning the experiments and logs directories of the last experiment
        mode=clear-experiment
        mode=clear-all # For clearing all the files from the experiments and logs directories/

"""

import os
import shutil
import argparse
import logging

import torch
import torch.utils.data as data
import numpy as np

from solver import Solver
from BrainMapperAE import BrainMapperAE3D, AutoEncoder3D
from utils.data_utils import get_datasets
from utils.settings import Settings
import utils.data_evaluation_utils as evaluations
from utils.common_utils import create_folder

# Set the default floating point tensor type to FloatTensor

torch.set_default_tensor_type(torch.FloatTensor)


def load_data(data_parameters, cross_domain_x2x_flag, cross_domain_y2y_flag):
    """Dataset Loader

    This function loads the training and validation datasets.

    Args:
        data_parameters (dict): Dictionary containing relevant information for the datafiles.
        cross_domain_x2x_flag (bool): Flag indicating if cross-domain training is occuring between the inputs
        cross_domain_y2y_flag (bool): Flag indicating if cross-domain training is occuring between the targets

    Returns:
        train_data (dataset object): Pytorch map-style dataset object, mapping indices to training data samples.
        validation_data (dataset object): Pytorch map-style dataset object, mapping indices to testing data samples.

    """
    print("Data is loading...")
    train_data, validation_data = get_datasets(data_parameters, cross_domain_x2x_flag, cross_domain_y2y_flag)
    print("Data has loaded!")
    print("Training dataset size is {}".format(len(train_data)))
    print("Validation dataset size is {}".format(len(validation_data)))

    return train_data, validation_data


def train(data_parameters, training_parameters, network_parameters, misc_parameters):
    """Training Function

    This function trains a given model using the provided training data.
    Currently, the data loaded is set to have multiple sub-processes.
    A high enough number of workers assures that CPU computations are efficiently managed, i.e. that the bottleneck is indeed the neural network's forward and backward operations on the GPU (and not data generation)
    Loader memory is also pinned, to speed up data transfer from CPU to GPU  by using the page-locked memory.
    Train data is also re-shuffled at each training epoch.

    Args:
        data_parameters (dict): Dictionary containing relevant information for the datafiles.

        training_parameters(dict): Dictionary containing relevant hyperparameters for training the network.
        training_parameters = {
            'training_batch_size': 5
            'validation_batch_size: 5
            'use_pre_trained': False
            'pre_trained_path': 'pre_trained/path'
            'experiment_name': 'experiment_name'
            'learning_rate': 1e-4
            'optimizer_beta': (0.9, 0.999)
            'optimizer_epsilon': 1e-8
            'optimizer_weigth_decay': 1e-5
            'number_of_epochs': 10
            'loss_log_period': 50
            'learning_rate_scheduler_step_size': 3
            'learning_rate_scheduler_gamma': 1e-1
            'use_last_checkpoint': True
            'final_model_output_file': 'path/to/model'
        }

        network_parameters (dict): Contains information relevant parameters

        misc_parameters (dict): Dictionary of aditional hyperparameters
        misc_parameters = {
            'save_model_directory': 'directory_name'
            'model_name': 'BrainMapper'
            'logs_directory': 'log-directory'
            'device': 1
            'experiments_directory': 'experiments-directory'
        }
    """


    def _load_pretrained_cross_domain(x2y_model, save_model_directory, experiment_name):
        """ Pretrained cross-domain loader

        This function loads the pretrained X2X and Y2Y autuencoders.
        After, it initializes the X2Y model's weights using the X2X encoder and teh Y2Y decoder weights.

        Args:
            x2y_model (class): Original x2y model initialised using the standard parameters.
            save_model_directory (str): Name of the directory where the model is saved
            experiment_name (str): Name of the experiment

        Returns:
            x2y_model (class): New x2y model with encoder and decoder paths weights reinitialised.
        """

        x2y_model_state_dict = x2y_model.state_dict()
        x2x_model_state_dict = torch.load(os.path.join(save_model_directory, experiment_name + '_x2x.pth.tar')).state_dict()
        y2y_model_state_dict = torch.load(os.path.join(save_model_directory, experiment_name + '_y2y.pth.tar')).state_dict()

        half_point = len(x2x_model_state_dict)//2 + 1

        counter = 1
        for key, _ in x2y_model_state_dict.items():
            if counter <= half_point:
                x2y_model_state_dict.update({key : x2x_model_state_dict[key]})
                counter+=1
            else:
                if key in y2y_model_state_dict:
                    x2y_model_state_dict.update({key : y2y_model_state_dict[key]})

        x2y_model.load_state_dict(x2y_model_state_dict)

        return x2y_model


    def _train_runner(data_parameters, 
                      training_parameters, 
                      network_parameters, 
                      misc_parameters,
                      optimizer = torch.optim.Adam,
                      loss_function = torch.nn.MSELoss(),
                      ):
        """Wrapper for the training operation

        This function wraps the training operation for the network

        Args:
            data_parameters (dict): Dictionary containing relevant information for the datafiles.
            training_parameters(dict): Dictionary containing relevant hyperparameters for training the network.
            network_parameters (dict): Contains information relevant parameters
            misc_parameters (dict): Dictionary of aditional hyperparameters

        """

        train_data, validation_data = load_data(data_parameters,
                                                cross_domain_x2x_flag = network_parameters['cross_domain_x2x_flag'], 
                                                cross_domain_y2y_flag = network_parameters['cross_domain_y2y_flag']
                                                )

        train_loader = data.DataLoader(
            dataset=train_data,
            batch_size=training_parameters['training_batch_size'],
            shuffle=True,
            pin_memory=True
        )

        validation_loader = data.DataLoader(
            dataset=validation_data,
            batch_size=training_parameters['validation_batch_size'],
            shuffle=False,
            pin_memory=True
        )

        if training_parameters['use_pre_trained']:
            BrainMapperModel = torch.load(training_parameters['pre_trained_path'])
        else:
            # BrainMapperModel = BrainMapperAE3D(network_parameters)
            BrainMapperModel = AutoEncoder3D(network_parameters) # temprorary change for testing encoder-decoder effective receptive field

        custom_weight_reset_flag = network_parameters['custom_weight_reset_flag']

        BrainMapperModel.reset_parameters(custom_weight_reset_flag)

        if network_parameters['cross_domain_x2y_flag'] == True:
            BrainMapperModel = _load_pretrained_cross_domain(x2y_model=BrainMapperModel, 
                                                             save_model_directory=misc_parameters['save_model_directory'], 
                                                             experiment_name=training_parameters['experiment_name']
                                                             )

        solver = Solver(model=BrainMapperModel,
                        device=misc_parameters['device'],
                        number_of_classes=network_parameters['number_of_classes'],
                        experiment_name=training_parameters['experiment_name'],
                        optimizer=optimizer,
                        optimizer_arguments={'lr': training_parameters['learning_rate'],
                                             'betas': training_parameters['optimizer_beta'],
                                             'eps': training_parameters['optimizer_epsilon'],
                                             'weight_decay': training_parameters['optimizer_weigth_decay']
                                             },
                        loss_function=loss_function,
                        model_name=training_parameters['experiment_name'],
                        number_epochs=training_parameters['number_of_epochs'],
                        loss_log_period=training_parameters['loss_log_period'],
                        learning_rate_scheduler_step_size=training_parameters[
                            'learning_rate_scheduler_step_size'],
                        learning_rate_scheduler_gamma=training_parameters['learning_rate_scheduler_gamma'],
                        use_last_checkpoint=training_parameters['use_last_checkpoint'],
                        experiment_directory=misc_parameters['experiments_directory'],
                        logs_directory=misc_parameters['logs_directory'],
                        checkpoint_directory=misc_parameters['checkpoint_directory'],
                        save_model_directory=misc_parameters['save_model_directory'],
                        crop_flag = data_parameters['crop_flag']
                        )

        validation_loss = solver.train(train_loader, validation_loader)

        del train_data, validation_data, train_loader, validation_loader, BrainMapperModel, solver, optimizer
        torch.cuda.empty_cache()

        return validation_loss


    if training_parameters['adam_w_flag'] == True:
        optimizer = torch.optim.AdamW
    else:
        optimizer = torch.optim.Adam

    loss_function = torch.nn.MSELoss()
    # loss_function=torch.nn.L1Loss()
    # loss_function=torch.nn.CosineEmbeddingLoss()

    if network_parameters['cross_domain_flag'] == False:
        _ = _train_runner(data_parameters, 
                          training_parameters, 
                          network_parameters, 
                          misc_parameters,
                          optimizer=optimizer,
                          loss_function=loss_function
                          )
    
    elif network_parameters['cross_domain_flag'] == True:
        if network_parameters['cross_domain_x2x_flag'] == True:

            training_parameters['experiment_name'] = training_parameters['experiment_name'] + '_x2x'
            data_parameters['target_data_train'] = data_parameters['input_data_train']
            data_parameters['target_data_validation'] = data_parameters['input_data_validation']

            # loss_function = torch.nn.L1Loss()

            _ = _train_runner(data_parameters, 
                            training_parameters, 
                            network_parameters, 
                            misc_parameters,
                            optimizer=optimizer,
                            loss_function=loss_function
                            )

        if network_parameters['cross_domain_y2y_flag'] == True:

            training_parameters['experiment_name'] = training_parameters['experiment_name'] + '_y2y'
            data_parameters['input_data_train'] = data_parameters['target_data_train']
            data_parameters['input_data_validation'] = data_parameters['target_data_validation']

            # loss_function = torch.nn.L1Loss()

            _ = _train_runner(data_parameters, 
                            training_parameters, 
                            network_parameters, 
                            misc_parameters,
                            optimizer=optimizer,
                            loss_function=loss_function
                            )

        if network_parameters['cross_domain_x2y_flag'] == True:

            _ = _train_runner(data_parameters, 
                            training_parameters, 
                            network_parameters, 
                            misc_parameters,
                            optimizer=optimizer,
                            loss_function=loss_function
                            )


def evaluate_mapping(mapping_evaluation_parameters):
    """Mapping Evaluator

    This function passes through the network an input and generates the rsfMRI outputs.

    Args:
        mapping_evaluation_parameters (dict): Dictionary of parameters useful during mapping evaluation.
        mapping_evaluation_parameters = {
            'trained_model_path': 'path/to/model'
            'data_directory': 'path/to/data'
            'data_list': 'path/to/datalist.txt/
            'prediction_output_path': 'directory-of-saved-predictions'
            'batch_size': 2
            'device': 0
            'exit_on_error': True
        }

    """
    trained_model_path = mapping_evaluation_parameters['trained_model_path']
    data_directory = mapping_evaluation_parameters['data_directory']
    mapping_data_file = mapping_evaluation_parameters['mapping_data_file']
    data_list = mapping_evaluation_parameters['data_list']
    prediction_output_path = mapping_evaluation_parameters['prediction_output_path']
    dmri_mean_mask_path = mapping_evaluation_parameters['dmri_mean_mask_path']
    rsfmri_mean_mask_path = mapping_evaluation_parameters['rsfmri_mean_mask_path']
    device = mapping_evaluation_parameters['device']
    exit_on_error = mapping_evaluation_parameters['exit_on_error']
    brain_mask_path = mapping_evaluation_parameters['brain_mask_path']
    regression_factors = mapping_evaluation_parameters['regression_factors']
    mean_regression_flag = mapping_evaluation_parameters['mean_regression_flag']
    mean_regression_all_flag = mapping_evaluation_parameters['mean_regression_all_flag']
    mean_subtraction_flag = mapping_evaluation_parameters['mean_subtraction_flag']
    scale_volumes_flag = mapping_evaluation_parameters['scale_volumes_flag']
    normalize_flag = mapping_evaluation_parameters['normalize_flag']
    minus_one_scaling_flag = mapping_evaluation_parameters['minus_one_scaling_flag']
    negative_flag = mapping_evaluation_parameters['negative_flag']
    outlier_flag = mapping_evaluation_parameters['outlier_flag']
    shrinkage_flag = mapping_evaluation_parameters['shrinkage_flag']
    hard_shrinkage_flag = mapping_evaluation_parameters['hard_shrinkage_flag']
    crop_flag = mapping_evaluation_parameters['crop_flag']

    evaluations.evaluate_mapping(trained_model_path,
                                 data_directory,
                                 mapping_data_file,
                                 data_list,
                                 prediction_output_path,
                                 brain_mask_path,
                                 dmri_mean_mask_path,
                                 rsfmri_mean_mask_path,
                                 regression_factors,
                                 mean_regression_flag,
                                 mean_regression_all_flag, 
                                 mean_subtraction_flag,
                                 scale_volumes_flag,
                                 normalize_flag,
                                 minus_one_scaling_flag,
                                 negative_flag, 
                                 outlier_flag,
                                 shrinkage_flag,
                                 hard_shrinkage_flag,
                                 crop_flag,
                                 device, 
                                 exit_on_error)


def delete_files(folder):
    """ Clear Folder Contents

    Function which clears contents (like experiments or logs)

    Args:
        folder (str): Name of folders whose conents is to be deleted

    """

    for object_name in os.listdir(folder):
        file_path = os.path.join(folder, object_name)
        try:
            if os.path.isfile(file_path):
                os.unlink(file_path)
            elif os.path.isdir(file_path):
                shutil.rmtree(file_path)
        except Exception as exception:
            print(exception)


if __name__ == '__main__':
    parser = argparse.ArgumentParser()
    parser.add_argument('--mode', '-m', required=True,
                        help='run mode, valid values are train or evaluate')
    parser.add_argument('--model_name', '-n', required=True,
                        help='model name, required for identifying the settings file modelName.ini & modelName_eval.ini')
    parser.add_argument('--use_last_checkpoint', '-c', required=False,
                        help='flag indicating if the last checkpoint should be used if 1; useful when wanting to time-limit jobs.')
    parser.add_argument('--number_of_epochs', '-e', required=False,
                        help='flag indicating how many epochs the network will train for; should be limited to ~3 hours or 2/3 epochs')

    arguments = parser.parse_args()

    settings_file_name = arguments.model_name + '.ini'
    evaluation_settings_file_name = arguments.model_name + '_eval.ini'

    settings = Settings(settings_file_name)
    data_parameters = settings['DATA']
    training_parameters = settings['TRAINING']
    network_parameters = settings['NETWORK']
    misc_parameters = settings['MISC']
    evaluation_parameters = settings['EVALUATION']

    if arguments.use_last_checkpoint == '1':
        training_parameters['use_last_checkpoint'] = True
    elif arguments.use_last_checkpoint == '0':
        training_parameters['use_last_checkpoint'] = False

    if arguments.number_of_epochs is not None:
        training_parameters['number_of_epochs'] = int(arguments.number_of_epochs)

    if arguments.mode == 'train':
        train(data_parameters, training_parameters,
              network_parameters, misc_parameters)

    elif arguments.mode == 'evaluate-mapping':
        logging.basicConfig(filename='evaluate-mapping-error.log')
        settings_evaluation = Settings(evaluation_settings_file_name)
        mapping_evaluation_parameters = settings_evaluation['MAPPING']
        evaluate_mapping(mapping_evaluation_parameters)
    elif arguments.mode == 'clear-experiments':
        shutil.rmtree(os.path.join(
            misc_parameters['experiments_directory'], training_parameters['experiment_name']))
        shutil.rmtree(os.path.join(
            misc_parameters['logs_directory'], training_parameters['experiment_name']))
        print('Cleared the current experiments and logs directory successfully!')
    elif arguments.mode == 'clear-everything':
        delete_files(misc_parameters['experiments_directory'])
        delete_files(misc_parameters['logs_directory'])
        print('Cleared the current experiments and logs directory successfully!')
    elif arguments.mode == 'train-and-evaluate-mapping':
        settings_evaluation = Settings(evaluation_settings_file_name)
        mapping_evaluation_parameters = settings_evaluation['MAPPING']
        train(data_parameters, training_parameters,
              network_parameters, misc_parameters)
        logging.basicConfig(filename='evaluate-mapping-error.log')
        evaluate_mapping(mapping_evaluation_parameters)
    else:
        raise ValueError(
            'Invalid mode value! Only supports: train, evaluate-score, evaluate-mapping, train-and-evaluate-mapping, clear-experiments and clear-everything')