Warning in sklean with using LinearPerSegmentModel and DateFlagsTransform

[Mr-Geekman]

🐛 Bug Report

Warning: Arrays of bytes/strings is being converted to decimal numbers if dtype='numeric'. This behavior is deprecated in 0.24 and will be removed in 1.1 (renaming of 0.26). Please convert your data to numeric values explicitly instead.

The problem is caused by category type of the results of DateFlagsTransform (even if they are boolean like in is_weekend).

This warning looks suspicious and something should be done about it.

Expected behavior

No warning.

How To Reproduce

from etna.datasets import generate_periodic_df
from etna.datasets import TSDataset
from etna.models import LinearPerSegmentModel
from etna.transforms import DateFlagsTransform 

classic_df = generate_periodic_df(periods=100, start_time="2020-01-01", n_segments=4, period=7, sigma=3)
df = TSDataset.to_dataset(df=classic_df)
ts = TSDataset(df, freq="D")
transforms = [DateFlagsTransform(is_weekend=True, day_number_in_week=False, day_number_in_month=False)]
ts.fit_transform(transforms)

model = LinearPerSegmentModel()
# here will be warning
model.fit(ts)
# here probably too
ts_future = ts.make_future(1)
model.forecast(ts_future)

Environment

No response

Additional context

No response

Checklist

• Bug appears at the latest library version
• Bug description added
• Steps to reproduce added
• Expected behavior added

[julia-shenshina]

need to be discussed

[julia-shenshina]

Add to_numeric method to *SklearnModels