How to Build Transformer#

Generate Transformer from Method#

[1]:

from __future__ import annotations

import numpy as np

from dtoolkit.transformer.factory import methodtf_factory

Generate a plus / minus constant transformer:

[2]:

def plus_constant(X: np.ndarray, constant: int | float) -> np.ndarray:
    """Plus constant to each element of ``X``"""

    return X + constant

def minus_constant(X: np.ndarray, constant: int | float) -> np.ndarray:
    """Minus constant to each element of ``X``"""

    return X - constant

PlusTF = methodtf_factory(plus_constant, minus_constant)

Use this transformer:

[3]:

a = np.array([1, 2, 3])
tf = PlusTF(constant=1).update_invargs(constant=1)

[4]:

tf.transform(a)

[4]:

array([2, 3, 4])

[5]:

tf.inverse_transform(a)

[5]:

array([0, 1, 2])

Build DataFrame Transformer#

Port numpy’s method to transformer.

[6]:

from dtoolkit.transformer import NumpyTF

class MyTF(NumpyTF):
    """Doc here"""

    transform_method = staticmethod("numpy's inner method")