Skip to content

glassbox.cleaner.encoders

OneHotEncoder and LabelEncoder for categorical feature encoding.

OneHotEncoder 

OneHotEncoder()

Bases: BaseTransformer

Encode categorical features as a one-hot numeric array.

Source code in glassbox/cleaner/encoders.py 
15
16
def __init__(self):
    self._categories: Dict[int, List[str]] = {}

fit 

fit(X)

Learn the categorical levels for encoding.

Parameters:

Name Type Description Default
X ndarray

Input array of shape (n_samples, n_features).

 required

Returns:

Type Description
Self

Fitted encoder instance.
Source code in glassbox/cleaner/encoders.py 
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
def fit(self, X: np.ndarray) -> Self:
    """
    Learn the categorical levels for encoding.

    Parameters
    ----------
    X : np.ndarray
        Input array of shape (n_samples, n_features).

    Returns
    -------
    Self
        Fitted encoder instance.
    """
    n_features = X.shape[1]
    for col_idx in range(n_features):
        col = X[:, col_idx]
        if np.issubdtype(col.dtype, np.number):
            col_clean = col[~np.isnan(col)]
        else:
            col_clean = np.array(
                [
                    x
                    for x in col
                    if x is not None and not (isinstance(x, float) and np.isnan(x))
                ]
            )

        uniques = np.unique(col_clean)
        self._categories[col_idx] = list(uniques)
    return self

transform 

transform(X)

Transform the dataset into a one-hot encoded representation.

Parameters:

Name Type Description Default
X ndarray

Input array of shape (n_samples, n_features).

 required

Returns:

Type Description
ndarray

Transformed array properly encoded.
Source code in glassbox/cleaner/encoders.py 
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
82
def transform(self, X: np.ndarray) -> np.ndarray:
    """
    Transform the dataset into a one-hot encoded representation.

    Parameters
    ----------
    X : np.ndarray
        Input array of shape (n_samples, n_features).

    Returns
    -------
    np.ndarray
        Transformed array properly encoded.
    """
    n_features = X.shape[1]
    out_cols = []
    for col_idx in range(n_features):
        col = X[:, col_idx]
        cats = self._categories.get(col_idx, [])
        if not cats:
            continue

        for cat in cats:
            if isinstance(cat, float) and np.isnan(cat):
                continue
            # For string objects, equal checking doesn't break, generates boolean mask
            mask = (col == cat).astype(float)
            out_cols.append(mask)

    if not out_cols:
        return np.empty((X.shape[0], 0))

    return np.column_stack(out_cols)

LabelEncoder 

LabelEncoder()

Bases: BaseTransformer

Encode target labels with value between 0 and n_classes-1.

Source code in glassbox/cleaner/encoders.py 
92
93
def __init__(self):
    self._mapping: Dict[str, int] = {}

fit 

fit(X)

Learn the vocabulary of the labels.

Parameters:

Name Type Description Default
X ndarray

Input array of shape (n_samples, n_features).

 required

Returns:

Type Description
Self

Fitted encoder instance.
Source code in glassbox/cleaner/encoders.py 
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
def fit(self, X: np.ndarray) -> Self:
    """
    Learn the vocabulary of the labels.

    Parameters
    ----------
    X : np.ndarray
        Input array of shape (n_samples, n_features).

    Returns
    -------
    Self
        Fitted encoder instance.
    """
    X_flat = X.ravel()
    if np.issubdtype(X_flat.dtype, np.number):
        X_clean = X_flat[~np.isnan(X_flat)]
    else:
        X_clean = np.array(
            [
                x
                for x in X_flat
                if x is not None and not (isinstance(x, float) and np.isnan(x))
            ]
        )

    uniques = np.unique(X_clean)
    self._mapping = {str(val): i for i, val in enumerate(uniques)}
    return self

transform 

transform(X)

Transform labels to normalized encoding.

Parameters:

Name Type Description Default
X ndarray

Input array of shape (n_samples, n_features).

 required

Returns:

Type Description
ndarray

Transformed array properly encoded.
Source code in glassbox/cleaner/encoders.py 
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
def transform(self, X: np.ndarray) -> np.ndarray:
    """
    Transform labels to normalized encoding.

    Parameters
    ----------
    X : np.ndarray
        Input array of shape (n_samples, n_features).

    Returns
    -------
    np.ndarray
        Transformed array properly encoded.
    """
    X_flat = X.ravel()
    res = np.zeros(X_flat.shape, dtype=float)

    for i, val in enumerate(X_flat):
        if (isinstance(val, float) and np.isnan(val)) or val is None:
            res[i] = np.nan
        else:
            res[i] = self._mapping.get(str(val), -1)

    return res.reshape(X.shape)