# License: BSD-3-Clause
import numpy as np
import pandas as pd
from sklearn.preprocessing import FunctionTransformer
from ..utils import check_Xs_y
from .. import deepmodule_installed, Tensor
if deepmodule_installed:
import torch
[docs]
class ObservedModIndicator(FunctionTransformer):
r"""
Binary indicator for observed modalities. Apply FunctionTransformer (from Scikit-learn) with
get_observed_mod_indicator as a function.
Note that this component typically should not be used in a vanilla Pipeline consisting of preprocessing and
an estimator.
See Also
--------
:class:`~imml.impute.MissingModIndicator`
:class:`~imml.impute.missing_mod_indicator`
:class:`~imml.impute.observed_mod_indicator`
Example
--------
>>> import numpy as np
>>> import pandas as pd
>>> from imml.impute import ObservedModIndicator
>>> from imml.ampute import Amputer
>>> Xs = [pd.DataFrame(np.random.default_rng(42).random((20, 10))) for i in range(3)]
>>> Xs = Amputer(p= 0.2, random_state=42).fit_transform(Xs)
>>> indicator = ObservedModIndicator()
>>> observed_mod = indicator.fit_transform()(Xs)
"""
def __init__(self):
super().__init__(get_observed_mod_indicator)
[docs]
def get_observed_mod_indicator(Xs : list, y = None):
r"""
Return a binary indicator for observed modalities.
Parameters
----------
Xs : list of array-likes objects
- Xs length: n_mods
- Xs[i] shape: (n_samples, n_features)
A list of different modalities.
y : Ignored
Not used, present here for API consistency by convention.
Returns
-------
transformed_X : array-likes objects, shape (n_samples, n_mods)
Binary indicator for observed modalities.
See Also
--------
:class:`~imml.impute.MissingModIndicator`
:class:`~imml.impute.ObservedModIndicator`
:class:`~imml.impute.missing_mod_indicator`
Example
--------
>>> import numpy as np
>>> import pandas as pd
>>> from imml.impute import get_observed_mod_indicator
>>> from imml.ampute import Amputer
>>> Xs = [pd.DataFrame(np.random.default_rng(42).random((20, 10))) for i in range(3)]
>>> Xs = Amputer(p= 0.2, random_state=42).fit_transform(Xs)
>>> observed_mod = get_observed_mod_indicator()(Xs)
"""
transformed_Xs = check_Xs_y(Xs, ensure_all_finite='allow-nan')
if isinstance(transformed_Xs[0], Tensor):
transformed_Xs = [X.numpy() for X in transformed_Xs]
transformed_X = np.vstack([pd.isna(X).all(1) for X in transformed_Xs]).T
transformed_X = np.logical_not(transformed_X)
if isinstance(Xs[0], pd.DataFrame):
transformed_X = pd.DataFrame(transformed_X, index=Xs[0].index)
elif isinstance(Xs[0], Tensor):
transformed_X = torch.from_numpy(transformed_X).bool()
return transformed_X