Discussion - Remove useless warning?

[MarieSacksick]

Problem Description

When using this simple piece of code, I have a warning:

from sklearn.datasets import make_classification
from skrub import tabular_pipeline

df, y = make_classification(random_state=42)
estimator = tabular_pipeline("classification")
estimator.fit(df, y)

The warning says:

UserWarning: Only pandas and polars DataFrames are supported, but input is a Numpy array. Please convert Numpy arrays to DataFrames before passing them to skrub transformers. Converting to pandas DataFrame with columns ['0', '1', …].
  warnings.warn(

I don't really understand the point of the warning, because it doesn't seem to be a problem at all.

Feature Description

What about removing the warning since having a numpy array as input doesn't seem to be a problem, at least in that use case with only one column, and doesn't call for a user action?
I'm also surprised by the fact that it seems to mention only the input y - why isn't df also a problem?

Alternative Solutions

Be more explicit about the consequences of automatic conversion. Otherwise as a user, I'm very tempted to turn off warning notifications.

Additional Context

No response

[revaarathore11]

@MarieSacksick can i work on this?

[MarieSacksick]

hello @revaarathore11, for now it's still a discussion, and not yet a proper issue to work on. I'd like to have the thoughts of other people before removing it.
From my current understanding right now, the message is between useless and misleading, because it states Only pandas and polars DataFrames are supported, and yet the code seems to process without problem, as if numpy arrays were sort of supported.

[rcap107]

I'm not entirely sure what is the reason behind the warning 🤔 it might be a leftover from some early version that did not support numpy arrays. I think we can remove the warning once we investigate why it's there in the first place.

[MarieSacksick]

@revaarathore11 would you want to explore that?
One way to do it is by checking when the line triggering the warning was added, find the skrub version, and either explore the code base from back then, or try to run the piece of code I wrote in the issue description and try to run it with this skrub version. If it is failing, then it means that the warning was there for a reason before, but it's not required anymore.