flatten

Flatten(features=None)

Bases: Transform

Flatten all time series in a DataFrame to individual features.

The given DataFrame's time series are converted into individual features, with each time step creating a new feature. This transform will change the order of the columns in the resulting dataset.

For example the dataset

series_data	A	B
{[0, 0], [1, 1], [2, 2]}	42	43
{[0, 3], [1, 4], [2, 5]}	44	45

gets flattened into the dataset

series_data_0	series_data_1	series_data_2	A	B
0	1	2	42	43
3	4	5	42	43

Initialize the flatten transform.

Parameters:

Name	Type	Description	Default
features	Iterable[str] | None	The features to flatten. If not provided or set to None, all possible features from the given dataframe will be flattened.	None

Source code in src/flowcean/transforms/flatten.py

def __init__(self, features: Iterable[str] | None = None) -> None:
    """Initialize the flatten transform.

    Args:
        features: The features to flatten. If not provided or set to None,
            all possible features from the given dataframe will be
            flattened.
    """
    self.features = features

FeatureLengthVaryError

Bases: Exception

Length of a feature varies over different rows.

NoTimeSeriesFeatureError

Bases: Exception

Feature is no time series.