Filters

eitprocessing.filters.butterworth_filters.LowPassFilter dataclass

LowPassFilter(
    *,
    filter_type: Literal["lowpass"] = "lowpass",
    cutoff_frequency: float | tuple[float],
    order: int,
    sample_frequency: float,
    ignore_max_order: InitVar[bool] = False
)

Low-pass Butterworth filter for filtering in the time domain.

LowPassFilter is a convenience class similar to ButterworthFilter, where the filter_type is set to "lowpass".

apply_filter

apply_filter(input_data: ArrayLike, axis: int = -1) -> ndarray

Apply the filter to the input data.

PARAMETER	DESCRIPTION
input_data	Data to be filtered. If the input data has more than one axis, the filter is applied to the last axis. TYPE: ArrayLike
axis	Data axis the filter should be applied to. This defaults to the last axis, assuming this to be the time axis of the input data. TYPE: int DEFAULT: -1

RETURNS	DESCRIPTION
ndarray	The filtered output with the same shape as the input data.

eitprocessing.filters.butterworth_filters.HighPassFilter dataclass

HighPassFilter(
    *,
    filter_type: Literal["highpass"] = "highpass",
    cutoff_frequency: float | tuple[float],
    order: int,
    sample_frequency: float,
    ignore_max_order: InitVar[bool] = False
)

High-pass Butterworth filter for filtering in the time domain.

HighPassFilter is a convenience class similar to ButterworthFilter, where the filter_type is set to "highpass".

apply_filter

apply_filter(input_data: ArrayLike, axis: int = -1) -> ndarray

Apply the filter to the input data.

PARAMETER	DESCRIPTION
input_data	Data to be filtered. If the input data has more than one axis, the filter is applied to the last axis. TYPE: ArrayLike
axis	Data axis the filter should be applied to. This defaults to the last axis, assuming this to be the time axis of the input data. TYPE: int DEFAULT: -1

RETURNS	DESCRIPTION
ndarray	The filtered output with the same shape as the input data.

eitprocessing.filters.butterworth_filters.BandStopFilter dataclass

BandStopFilter(
    *,
    filter_type: Literal["bandstop"] = "bandstop",
    cutoff_frequency: float | tuple[float],
    order: int,
    sample_frequency: float,
    ignore_max_order: InitVar[bool] = False
)

Band-stop Butterworth filter for filtering in the time domain.

BandStopFilter is a convenience class similar to ButterworthFilter, where the filter_type is set to "bandstop".

apply_filter

apply_filter(input_data: ArrayLike, axis: int = -1) -> ndarray

Apply the filter to the input data.

PARAMETER	DESCRIPTION
input_data	Data to be filtered. If the input data has more than one axis, the filter is applied to the last axis. TYPE: ArrayLike
axis	Data axis the filter should be applied to. This defaults to the last axis, assuming this to be the time axis of the input data. TYPE: int DEFAULT: -1

RETURNS	DESCRIPTION
ndarray	The filtered output with the same shape as the input data.

eitprocessing.filters.butterworth_filters.BandPassFilter dataclass

BandPassFilter(
    *,
    filter_type: Literal["bandpass"] = "bandpass",
    cutoff_frequency: float | tuple[float],
    order: int,
    sample_frequency: float,
    ignore_max_order: InitVar[bool] = False
)

Band-pass Butterworth filter for filtering in the time domain.

BandPassFilter is a convenience class similar to ButterworthFilter, where the filter_type is set to "bandpass".

apply_filter

apply_filter(input_data: ArrayLike, axis: int = -1) -> ndarray

Apply the filter to the input data.

PARAMETER	DESCRIPTION
input_data	Data to be filtered. If the input data has more than one axis, the filter is applied to the last axis. TYPE: ArrayLike
axis	Data axis the filter should be applied to. This defaults to the last axis, assuming this to be the time axis of the input data. TYPE: int DEFAULT: -1

RETURNS	DESCRIPTION
ndarray	The filtered output with the same shape as the input data.

eitprocessing.filters.butterworth_filters.ButterworthFilter dataclass

ButterworthFilter(
    *,
    filter_type: Literal["lowpass", "highpass", "bandpass", "bandstop"],
    cutoff_frequency: float | tuple[float],
    order: int,
    sample_frequency: float,
    ignore_max_order: InitVar[bool] = False
)

Butterworth filter for filtering in the time domain.

Generates a low-pass, high-pass, band-pass or band-stop digital Butterworth filter of order order. Filters are created using cascaded second-order sections representation, providing better stability compared to the traditionally used transfer function (numerator/denominator or b/a representation).

The apply_filter() method applies the filter to the provided data using forward-backward filtering. This minimizes the phase shift, and effectively doubles the order of the filter.

ButterworthFilter is a wrapper of the scipy.butter() and scipy.filtfilt() functions: - https://docs.scipy.org/doc/scipy-1.10.1/reference/generated/scipy.signal.butter.html - https://docs.scipy.org/doc/scipy-1.10.1/reference/generated/scipy.signal.filtfilt.html

PARAMETER	DESCRIPTION
filter_type	The type of filter to create: a low pass, high pass, band pass or band stop filter. TYPE: Literal['lowpass', 'highpass', 'bandpass', 'bandstop']
cutoff_frequency	Cutoff frequency or frequencies (in Hz). For low pass or high pass filters, cutoff_frequency is a scalar. For band pass or band stop filters, cutoff_frequency is a sequence containing two frequencies. TYPE: float | tuple[float]
order	Order of the filter. The effective order size is twice the given order, due to forward-backward filtering. Higher orders improve the effectiveness of a filter, but can result in unstable or incorrect filtering. TYPE: int
sample_frequency	The sample frequency of the data to be filtered (in Hz). TYPE: float
ignore_max_order	Whether to raise an exception if the order is larger than the maximum of 10. Defaults to False. TYPE: InitVar[bool] DEFAULT: False

Examples:

>>> t = np.arange(0, 100, 0.1)
>>> signal = np.sin(t) + 0.1 * np.sin(10 * t)
>>> lowpass_filter = ButterworthFilter(
...     filter_type='lowpass',
...     cutoff_frequenct=45,
...     order=4,
...     sample_frequency=250
... )
>>> filtered_signal = lowpass_filter.apply_filter(signal)

apply_filter

apply_filter(input_data: ArrayLike, axis: int = -1) -> ndarray

Apply the filter to the input data.

PARAMETER	DESCRIPTION
input_data	Data to be filtered. If the input data has more than one axis, the filter is applied to the last axis. TYPE: ArrayLike
axis	Data axis the filter should be applied to. This defaults to the last axis, assuming this to be the time axis of the input data. TYPE: int DEFAULT: -1

RETURNS	DESCRIPTION
ndarray	The filtered output with the same shape as the input data.