[TO_REVIEW] Bug fix when None in make_da_pipeline

[antoinecollas]

Enhancement: Add compatibility with None in skada pipelines

This PR introduces support for using None within skada pipelines, similar to how scikit-learn handles it. For example:

pipe = make_pipeline(
        StandardScaler(),
        None,  # Skip step
        PCA(),
        LogisticRegression(),
)

Now, skada pipelines can also include None as a step:

pipe = make_da_pipeline(
        StandardScaler(),
        None,  # Skip step
        PCA(),
        SubspaceAlignmentAdapter(n_components=2),
        LogisticRegression(),
)

[rflamary]

LGTM, I would maybe test with and without the None to check that we have the same output.