Module monk.system.eda.eda

Functions

def class_imbalance(system_dict, show_img, save_img)

Find class imbalance

Args

system_dict : dict

System dictionary storing experiment state and set variables

show_img : bool

If True, displays bar graph for images per class

save_img : bool

If True, saves bar graph for images per class

Returns

list

List of classes

list

List of number of images in every class

Expand source code

def class_imbalance(system_dict, show_img, save_img):
    '''
    Find class imbalance

    Args:
        system_dict (dict): System dictionary storing experiment state and set variables
        show_img (bool): If True, displays bar graph for images per class 
        save_img (bool): If True, saves bar graph for images per class

    Returns:
        list: List of classes
        list: List of number of images in every class
    '''

    dataset_type = system_dict["dataset"]["dataset_type"];
    dataset_train_path = system_dict["dataset"]["train_path"];
    dataset_val_path = system_dict["dataset"]["val_path"];
    delimiter = system_dict["dataset"]["params"]["delimiter"];
    log_dir = system_dict["log_dir"];

    if("csv" in dataset_type):
        csv_train = system_dict["dataset"]["csv_train"];
        csv_val = system_dict["dataset"]["csv_val"];

    if(dataset_type == "train"):
        classes_folder, classes_folder_strength = populate_from_folder_train(dataset_train_path);                   
    elif(dataset_type == "train-val"):
        classes_folder, classes_folder_strength = populate_from_folder_trainval(dataset_train_path, dataset_val_path);

    elif(dataset_type == "csv_train"):
        classes_folder, classes_folder_strength = populate_from_csv_train(csv_train, delimiter);

    elif(dataset_type == "csv_train-val"):
        classes_folder, classes_folder_strength = populate_from_csv_trainval(csv_train, csv_val, delimiter);

    create_plot(classes_folder, classes_folder_strength, log_dir, show_img, save_img);

    return classes_folder, classes_folder_strength;

def corrupted_missing_images(system_dict, check_missing, check_corrupt)

Find corrupt and missing images in dataset

Args

system_dict : dict

System dictionary storing experiment state and set variables

check_missing : bool

If True, checks for missing images in csv type dataset

check_corrupt : bool

If True, checks for corrupted images in foldered and csv dataset

Returns

list

List of images missing from training set

list

List of images missing from validation set

list

List of images corrupted in training set

list

List of images corrupted in validation set

Expand source code

def corrupted_missing_images(system_dict, check_missing, check_corrupt):
    '''
    Find corrupt and missing images in dataset

    Args:
        system_dict (dict): System dictionary storing experiment state and set variables
        check_missing (bool): If True, checks for missing images in csv type dataset
        check_corrupt (bool): If True, checks for corrupted images in foldered and csv dataset

    Returns:
        list: List of images missing from training set
        list: List of images missing from validation set
        list: List of images corrupted in training set
        list: List of images corrupted in validation set
    '''
    dataset_type = system_dict["dataset"]["dataset_type"];
    dataset_train_path = system_dict["dataset"]["train_path"];
    dataset_val_path = system_dict["dataset"]["val_path"];

    if("csv" in dataset_type):
        csv_train = system_dict["dataset"]["csv_train"];
        csv_val = system_dict["dataset"]["csv_val"];
        delimiter = system_dict["dataset"]["params"]["delimiter"];

    missing_images_train = None;
    missing_images_val = None;
    corrupt_images_train = None;
    corrupt_images_val = None;



    if(dataset_type == "train"):
        if(check_missing):
            x=0;
        if(check_corrupt):
            corrupt_images_train = populate_corrupt_from_foldered(dataset_train_path, verbose=system_dict["verbose"]);
    elif(dataset_type == "train-val"):
        if(check_missing):
            x=0;
        if(check_corrupt):
            corrupt_images_train = populate_corrupt_from_foldered(dataset_train_path, verbose=system_dict["verbose"])
            corrupt_images_val = populate_corrupt_from_foldered(dataset_val_path, verbose=system_dict["verbose"])
    elif(dataset_type == "csv_train"):
        if(check_missing):
            missing_images_train = populate_missing(csv_train, dataset_train_path, delimiter);
        if(check_corrupt):
            corrupt_images_train = populate_corrupt_from_csv(csv_train, dataset_train_path, delimiter, verbose=system_dict["verbose"]);
    elif(dataset_type == "csv_train-val"):
        if(check_missing):
            missing_images_train = populate_missing(csv_train, dataset_train_path, delimiter);
            missing_images_val = populate_missing(csv_val, dataset_val_path, delimiter);
        if(check_corrupt):
            corrupt_images_train = populate_corrupt_from_csv(csv_train, dataset_train_path, delimiter, verbose=system_dict["verbose"]);
            corrupt_images_val = populate_corrupt_from_csv(csv_val, dataset_val_path, delimiter, verbose=system_dict["verbose"]);

    return missing_images_train, missing_images_val, corrupt_images_train, corrupt_images_val;