Bug: KeyError when "X" column is missing in DataFrame in windowing_df function

[gkontogiannhs]

The windowing_df function assumes that the input DataFrame contains a column named "X", but this assumption is not checked. If the "X" column is missing or incorrectly structured, the expression row["X"] will raise a KeyError.

Additionally, even if the "X" column exists, the function processes it row by row, treating each row["X"] as a single scalar value. This is problematic because it results in feeding a scalar value into the rolling_window_df function, which expects an iterable. This leads to unexpected behavior, particularly since windowing_df is designed to iterate through rows of the DataFrame and apply a sliding window algorithm, an approach that is unconventional for this library's typical standards (column-wise iteration treating each column as a separate dimension).