elasticai.hw_measurements.process.data

Module Contents

Classes

MetricCalculator

Functions

window_method	Generating window for smoothing transformation method.
do_fft	Performing the Discrete Fast Fourier Transformation.
calculate_total_harmonics_distortion	Calculating the Total Harmonics Distortion (THD) of spectral input
calculate_total_harmonics_distortion_from_transient	Calculating the Total Harmonics Distortion (THD) from transient input Args: signal: Array with frequency values for spectral analysis fs: Sampling rate [Hz] N_harmonics: Number of used harmonics for calculating THD Return: THD value (in dB)

API

elasticai.hw_measurements.process.data.window_method(window_size: int, method: str = 'hamming') → numpy.ndarray

Generating window for smoothing transformation method.

Parameters:

• window_size – Integer number with size of the window

• method – Selection of window method [‘’: None, ‘Hamming’, ‘guassian’, ‘bartlett’, ‘blackman’]

Returns:

Numpy array with window

elasticai.hw_measurements.process.data.do_fft(y: numpy.ndarray, fs: float, method_window: str = '') → elasticai.hw_measurements.TransformSpectrum

Performing the Discrete Fast Fourier Transformation.

Parameters:

• y – Transient input signal

• fs – Sampling rate [Hz]

• method_window – Selected window [‘’: None, ‘Hamming’, ‘guassian’, ‘bartlett’, ‘blackman’]

Returns:

Dataclass TransformSpectrum with spectrum data

elasticai.hw_measurements.process.data.calculate_total_harmonics_distortion(data: elasticai.hw_measurements.TransformSpectrum, N_harmonics: int = 4) → float

Calculating the Total Harmonics Distortion (THD) of spectral input

Parameters:

• data – Dataclass TransformSpectrum

• N_harmonics – Number of used harmonics for calculating THD Return: THD value (in dB) and corresponding frequency positions of peaks

elasticai.hw_measurements.process.data.calculate_total_harmonics_distortion_from_transient(signal: numpy.ndarray, fs: float, N_harmonics: int = 4) → float

Calculating the Total Harmonics Distortion (THD) from transient input Args: signal: Array with frequency values for spectral analysis fs: Sampling rate [Hz] N_harmonics: Number of used harmonics for calculating THD Return: THD value (in dB)

class elasticai.hw_measurements.process.data.MetricCalculator

Bases: elasticai.hw_measurements.process.common.ProcessCommon

Initialization

Class with constructors for processing the measurement results for extracting device-specific metrics

calculate_lsb_mean(stim_input: numpy.ndarray, daq_output: numpy.ndarray) → float

Function for calculating the mean Least Significant Bit (LSB)

Parameters:

• stim_input – Numpy array with stimulus input

• daq_output – Numpy array with DAQ output

Returns:

Float with LSB

static calculate_gain_from_transfer(stim_input: numpy.ndarray, src_output: numpy.ndarray) → numpy.ndarray

Function for extracting the gain of an electrical circuit using the transfer function test

Parameters:

• stim_input – Numpy array with stimulus input

• daq_output – Numpy array with DAQ output (should have same unit like stim_input)

Returns:

Numpy array with gain value

static calculate_lsb(stim_input: numpy.ndarray, daq_output: numpy.ndarray) → numpy.array

Function for calculating the mean Least Significant Bit (LSB)

Parameters:

• stim_input – Numpy array with stimulus input

• daq_output – Numpy array with DAQ output

Returns:

Float with LSB

calculate_dnl(stim_input: numpy.ndarray, daq_output: numpy.ndarray) → numpy.ndarray

Calculating the Differential Non-Linearity (DNL) of a transfer function from DAC/ADC

Parameters:

• stim_input – Numpy array with stimulus input

• daq_output – Numpy array with DAQ output

Returns:

Numpy array with DNL

calculate_inl(stim_input: numpy.ndarray, daq_output: numpy.ndarray) → numpy.ndarray

Calculating the Integral Non-Linearity (INL) of a transfer function from DAC/ADC

Parameters:

• stim_input – Numpy array with stimulus input

• daq_output – Numpy array with DAQ output

Returns:

Numpy array with INL

static calculate_error_mbe(y_pred: numpy.ndarray | float, y_true: numpy.ndarray | float) → float

Calculating the distance-based metric with mean bias error

Parm y_pred:

Numpy array or float value from prediction

Parameters:

y_true – Numpy array or float value from true label

Returns:

Float value with error

static calculate_error_mae(y_pred: numpy.ndarray | float, y_true: numpy.ndarray | float) → float

Calculating the distance-based metric with mean absolute error

Parameters:

• y_pred – Numpy array or float value from prediction

• y_true – Numpy array or float value from true label

Returns:

Float value with error

static calculate_error_mse(y_pred: numpy.ndarray | float, y_true: numpy.ndarray | float) → float

Calculating the distance-based metric with mean squared error

Parameters:

• y_pred – Numpy array or float value from prediction

• y_true – Numpy array or float value from true label

Returns:

Float value with error

static calculate_error_mape(y_pred: numpy.ndarray | float, y_true: numpy.ndarray | float) → float

Calculating the distance-based metric with mean absolute percentage error

Parameters:

• y_pred – Numpy array or float value from prediction

• y_true – Numpy array or float value from true label

Returns:

Float value with error

static calculate_total_harmonics_distortion(signal: numpy.ndarray, fs: float, N_harmonics: int) → float

Calculating the Total Harmonics Distortion (THD) of transient input

Parameters:

• signal – Numpy array with transient signal to extract spectral analysis

• fs – Applied sampling rate [Hz]

• N_harmonics – Number of used harmonics for calculating THD

Returns:

THD value (in dB)

static calculate_cosine_similarity(y_pred: numpy.ndarray, y_true: numpy.ndarray) → float

Calculating the Cosine Similarity of two different inputs (same size)

Parameters:

• y_pred – Numpy array or float value from prediction

• y_true – Numpy array or float value from true label

Returns:

Float value with error