DSLR/data_preprocessing.py at main · plagache/DSLR · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
import pandas

from f_statistics import percentile
from variables import labels_column


def create_dataframe(csv_string):
    csv_dataset = pandas.read_csv(csv_string)
    dataset = pandas.DataFrame(csv_dataset)
    dataset.sort_index(axis=1, inplace=True)
    return dataset


def cleanup_nan(dataframe):
    cleaned_series = []
    for _, serie in dataframe.items():
        serie.dropna(inplace=True, ignore_index=True)
        cleaned_series.append(serie)
    return pandas.concat(cleaned_series, axis=1)


def split_by_classes(dataframe):
    classes = create_classes(dataframe)
    data_classes = []
    for class_name in classes:
        data_classes.append(dataframe.loc[dataframe[labels_column] == class_name])
    return data_classes


def get_selected_features(dataset, features_to_select):
    return dataset[features_to_select]


def select_numerical_features(dataset):
    return dataset.select_dtypes(include=["float64"])


def create_labels(dataset, classes):
    labels = pandas.DataFrame()
    for class_name in classes:
        labels[class_name] = dataset[labels_column].map(lambda x: 1.0 if x == class_name else 0.0)
    return labels


def create_classes(dataset):
    classes = dataset[labels_column].unique()
    classes.sort()
    return classes


def split_dataframe(dataframe, test_percent: float):
    test_sample = dataframe.groupby(labels_column).sample(frac=test_percent)
    train_sample = dataframe.drop(test_sample.index)
    return test_sample, train_sample


def numerization(dataset):
    dataset["Birthday"] = pandas.to_datetime(dataset["Birthday"]).astype(int)
    numerized_dataset = dataset.replace(regex={"Right": 1, "Left": -1})
    return numerized_dataset


def get_quartiles(dataframe):
    quartiles = []
    for _, data in dataframe.items():
        first = percentile(data, 0.25)
        second = percentile(data, 0.5)
        third = percentile(data, 0.75)
        quartiles.append((data.name, first, second, third))
    return quartiles


def robust_scale(dataframe: pandas.DataFrame, quartiles):
    """
    Replace nan with median value
    Then scale
    """
    ret = pandas.DataFrame()
    for (name, data), (courses, first, second, third) in zip(dataframe.items(), quartiles):
        ret[name] = data.fillna(second).map(lambda x: (x - second) / (third - first))
    return ret