.predict in pipeline

[PolarsDude]

Hi

Consider the following pipeline (from examples)

import numpy as np
import pandas as pd

from sklearn.pipeline import FeatureUnion, Pipeline
from sklearn.linear_model import LinearRegression, Ridge

from sklego.meta import EstimatorTransformer
from sklego.preprocessing import ColumnSelector

np.random.seed(42)
n = 1000
X = np.random.uniform(0, 1, (n, 2))
y = X.sum(axis=1) + np.random.uniform(0, 1, (n,))
df = pd.DataFrame({"x1": X[:, 0], "x2": X[:, 1], "y": y})

pipeline = Pipeline([
    ("grab_columns", ColumnSelector(["x1", "x2"])),
    ("ml_features", FeatureUnion([
        ("model_1",  EstimatorTransformer(LinearRegression())),
        ("model_2",  EstimatorTransformer(Ridge()))
    ]))
])

pipeline.fit(df, y)

I want to predict with this pipeline with new data, for instance pipeline.predict(X_test), but I get the follow error message: AttributeError: 'EstimatorTransformer' object has no attribute 'predict'. Is there any way to use predict in pipeline? This would help alot.

[FBruzzesi]

Hey @PolarsDude thanks for the question.
EstimatorTransformer turns any estimator to a transformer component, therefore you can call pipeline.transform after fitting.
If you want to use .predict, then this might not be the component you are looking for or you need to follow it with an estimator.